chore(release): cut 0.6.1

* test(app): consolidate Playwright journeys

* test(app): stabilize consolidated Playwright flows

* test(app): make streaming dependencies explicit

* test(app): await profile feature persistence

* test(app): sample only actual workspace tabs

.github/PULL_REQUEST_TEMPLATE.md

@@ -18,7 +18,7 @@ Closes #
 
 ### Reasoning
 
-<!-- A short description of the reasoning of this change in your own words. What was wrong, and how do you hope this PR fixes it. If you can't explain this briefly, the PR is probably too big. This description must be grounded in real user flows and framed as value provider to Paseo users. -->
+<!-- A short description of the reasoning of this change in your own words. What was wrong, and how do you hope this PR fixes it. If you can't explain this briefly, the PR is probably too big. This description must be grounded in real user flows and framed as value provided to Paseo users. -->
 
 ### Goals
 

.github/workflows/ci.yml

@@ -348,12 +348,12 @@ jobs:
         timeout-minutes: 10
         run: npx playwright install chromium
 
-      - name: Build app dependencies
-        run: npm run build:app-deps
-
       - name: Build server stack
         run: npm run build:server
 
+      - name: Build native app dependency
+        run: npm run build --workspace=@getpaseo/expo-two-way-audio
+
       - name: Install agent CLIs for provider tests
         run: node scripts/npm-retry.mjs install -g @anthropic-ai/claude-code @openai/codex@0.105.0 opencode-ai
 

.github/workflows/docker.yml

@@ -18,6 +18,7 @@ on:
       - "packages/client/**"
       - "packages/expo-two-way-audio/**"
       - "packages/highlight/**"
+      - "packages/plugin/**"
       - "packages/protocol/**"
       - "packages/relay/**"
       - "packages/server/**"

.github/workflows/nix-update-hash.yml

@@ -10,6 +10,7 @@ on:
       - "package.json"
       - "package-lock.json"
       - "packages/highlight/**"
+      - "packages/plugin/**"
       - "packages/server/**"
       - "packages/relay/**"
       - "packages/cli/**"

.github/workflows/nix.yml

@@ -12,6 +12,7 @@ on:
       - "packages/app/**"
       - "packages/expo-two-way-audio/**"
       - "packages/highlight/**"
+      - "packages/plugin/**"
       - "packages/protocol/**"
       - "packages/client/**"
       - "packages/server/**"

.oxlintrc.json

@@ -94,6 +94,11 @@
           "error",
           {
             "paths": [
+              {
+                "name": "react-native",
+                "importNames": ["TextInput"],
+                "message": "Use EditingTextInput from @/components/ui/text-input. Raw TextInput bypasses editing-surface authority and IME safety."
+              },
               {
                 "name": "@tanstack/react-query",
                 "importNames": ["useQuery", "useInfiniteQuery", "useQueries"],
@@ -120,6 +125,11 @@
           "error",
           {
             "paths": [
+              {
+                "name": "react-native",
+                "importNames": ["TextInput"],
+                "message": "Use EditingTextInput from @/components/ui/text-input. Raw TextInput bypasses editing-surface authority and IME safety."
+              },
               {
                 "name": "react-native-unistyles",
                 "importNames": ["useUnistyles"],
@@ -167,6 +177,11 @@
           "error",
           {
             "paths": [
+              {
+                "name": "react-native",
+                "importNames": ["TextInput"],
+                "message": "Use EditingTextInput from @/components/ui/text-input. Raw TextInput bypasses editing-surface authority and IME safety."
+              },
               {
                 "name": "react-native-unistyles",
                 "importNames": ["useUnistyles"],
@@ -192,11 +207,11 @@
         "packages/app/src/desktop/browser/pane/index.tsx",
         "packages/app/src/desktop/browser/pane/index.web.tsx",
         "packages/app/src/components/command-center.tsx",
+        "packages/app/src/components/compact-explorer-sidebar.tsx",
         "packages/app/src/components/context-window-meter.tsx",
         "packages/app/src/components/dictation-controls.tsx",
         "packages/app/src/components/download-toast.tsx",
         "packages/app/src/components/draggable-list.native.tsx",
-        "packages/app/src/components/explorer-sidebar.tsx",
         "packages/app/src/components/headers/back-header.tsx",
         "packages/app/src/components/headers/menu-header.tsx",
         "packages/app/src/components/headers/screen-header.tsx",
@@ -254,6 +269,11 @@
           "error",
           {
             "paths": [
+              {
+                "name": "react-native",
+                "importNames": ["TextInput"],
+                "message": "Use EditingTextInput from @/components/ui/text-input. Raw TextInput bypasses editing-surface authority and IME safety."
+              },
               {
                 "name": "@tanstack/react-query",
                 "importNames": ["useQuery", "useInfiniteQuery", "useQueries"],
@@ -274,6 +294,26 @@
         "packages/app/src/desktop/components/pair-device-section.tsx",
         "packages/app/src/screens/new-workspace-screen.tsx"
       ],
+      "rules": {
+        "no-restricted-imports": [
+          "error",
+          {
+            "paths": [
+              {
+                "name": "react-native",
+                "importNames": ["TextInput"],
+                "message": "Use EditingTextInput from @/components/ui/text-input. Raw TextInput bypasses editing-surface authority and IME safety."
+              }
+            ]
+          }
+        ]
+      }
+    },
+    {
+      "files": [
+        "packages/app/src/components/ui/text-input/**/*.{ts,tsx}",
+        "packages/app/src/agent-stream/scroll-keyboard-dismiss/use-scroll-keyboard-dismiss.ts"
+      ],
       "rules": {
         "no-restricted-imports": "off"
       }

CHANGELOG.md

@@ -1,5 +1,199 @@
 # Changelog
 
+## 0.6.1 - 2026-08-25
+
+### Improved
+
+- Command Center now matches partial multi-word queries in any order and ranks visible results by relevance ([#2971](https://github.com/getpaseo/paseo/pull/2971) by [@cleiter](https://github.com/cleiter))
+
+### Fixed
+
+- Fixed upgrades placing persisted Side panel tabs inside Explorer instead of an ordinary side pane ([#3861](https://github.com/getpaseo/paseo/pull/3861))
+- Fixed Escape leaving Settings open on desktop and web ([#2828](https://github.com/getpaseo/paseo/pull/2828) by [@mcowger](https://github.com/mcowger))
+- Fixed branch switches hiding the current pull request or arming auto-archive from an already-merged pull request ([#3799](https://github.com/getpaseo/paseo/pull/3799))
+
+## 0.6.0 - 2026-08-25
+
+### Changed
+
+- Restored the dedicated Explorer sidebar with Files and Changes by default, replacing the empty user-directed Side panel introduced in 0.5.0 ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+- Changed content opened to the side to use an ordinary closable workspace pane instead of the Explorer sidebar ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+- Restored the resizable combined Explorer dock on tablets ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+
+### Added
+
+- Added per-action settings for opening files, changes, subagents, and pull requests in the main panel or on the side ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+- Added New tab and drag support for compatible files, diffs, agents, terminals, pull requests, and plugin panels in Explorer ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+- Added standalone Diff tabs while keeping an optional persisted inline-diff mode in Changes ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+- Added custom Windows and Linux window controls ([#3826](https://github.com/getpaseo/paseo/pull/3826))
+
+### Fixed
+
+- Fixed OpenCode turns failing when the first event-stream connection stalls during startup ([#3814](https://github.com/getpaseo/paseo/pull/3814))
+- Fixed OpenCode prompts starting before the provider reports a ready connection ([#3821](https://github.com/getpaseo/paseo/pull/3821))
+
+## 0.5.2 - 2026-08-24
+
+### Fixed
+
+- Fixed workspaces crashing on Android tablets and foldables in non-compact layouts ([#3760](https://github.com/getpaseo/paseo/pull/3760) by [@jeid64](https://github.com/jeid64))
+- Fixed app settings resetting when builds sharing a profile recognize different settings ([#3787](https://github.com/getpaseo/paseo/pull/3787))
+
+## 0.5.1 - 2026-08-23
+
+### Fixed
+
+- Fixed multiline composers stopping resizing after workspace and agent transitions on iOS and Android ([#3740](https://github.com/getpaseo/paseo/pull/3740))
+
+## 0.5.0 - 2026-08-23
+
+### Added
+
+- Added experimental trusted local plugins ([#3222](https://github.com/getpaseo/paseo/pull/3222), [#3446](https://github.com/getpaseo/paseo/pull/3446), [#3465](https://github.com/getpaseo/paseo/pull/3465))
+- Added plugin-contributed application themes to Settings → Appearance ([#3602](https://github.com/getpaseo/paseo/pull/3602) by [@ragokan](https://github.com/ragokan))
+- Added a user-directed Side panel with independent tabs ([#3287](https://github.com/getpaseo/paseo/pull/3287), [#3605](https://github.com/getpaseo/paseo/pull/3605))
+- Added browser-style New tabs with an anchored pane-local chooser ([#3715](https://github.com/getpaseo/paseo/pull/3715), [#3735](https://github.com/getpaseo/paseo/pull/3735))
+- Added active-turn steering for Claude, Codex, and OpenCode ([#3394](https://github.com/getpaseo/paseo/pull/3394), [#3580](https://github.com/getpaseo/paseo/pull/3580) by [@mcowger](https://github.com/mcowger))
+- Added workspace labels for sidebar organization and filtering ([#3510](https://github.com/getpaseo/paseo/pull/3510) by [@cleiter](https://github.com/cleiter))
+- Added live workspace change counts above active agent composers ([#3682](https://github.com/getpaseo/paseo/pull/3682))
+- Added workspace panel, tab, and pane actions to the Command Center ([#3685](https://github.com/getpaseo/paseo/pull/3685))
+- Added `paseo project` to create, list, rename, and delete projects from the terminal ([#3460](https://github.com/getpaseo/paseo/pull/3460))
+- Added `paseo reload` to apply runtime-safe `config.json` changes without restarting the daemon ([#3365](https://github.com/getpaseo/paseo/pull/3365))
+- Added Paseo skill management to Host → Agents, including remote hosts ([#3451](https://github.com/getpaseo/paseo/pull/3451))
+- Added project filtering to the sidebar display menu ([#3563](https://github.com/getpaseo/paseo/pull/3563) by [@cleiter](https://github.com/cleiter))
+- Added optional branch and project names on workspace rows ([#3445](https://github.com/getpaseo/paseo/pull/3445))
+- Added drag-and-drop reordering for pinned workspaces ([#3341](https://github.com/getpaseo/paseo/pull/3341))
+- Added one-click agent profile creation from the model chooser ([#3533](https://github.com/getpaseo/paseo/pull/3533))
+- Added Android Studio to Open in editor ([#3531](https://github.com/getpaseo/paseo/pull/3531), [#3614](https://github.com/getpaseo/paseo/pull/3614) by [@caikovsky](https://github.com/caikovsky))
+- Added MiniMax Code to the one-click ACP provider catalog ([#3457](https://github.com/getpaseo/paseo/pull/3457) by [@hetaoBackend](https://github.com/hetaoBackend))
+- Added Nix syntax highlighting ([#3110](https://github.com/getpaseo/paseo/pull/3110) by [@Strainy](https://github.com/Strainy))
+- Added Svelte syntax highlighting ([#3487](https://github.com/getpaseo/paseo/pull/3487), [#3534](https://github.com/getpaseo/paseo/pull/3534) by [@fiorelorenzo](https://github.com/fiorelorenzo))
+- Added guided Hub setup that connects a daemon and deploys a compatible starter workflow ([#3651](https://github.com/getpaseo/paseo/pull/3651), [#3657](https://github.com/getpaseo/paseo/pull/3657), [#3677](https://github.com/getpaseo/paseo/pull/3677))
+
+### Improved
+
+- Restored cached projects, workspaces, agents, and timelines immediately while hosts reconnect ([#3259](https://github.com/getpaseo/paseo/pull/3259), [#3329](https://github.com/getpaseo/paseo/pull/3329))
+- Reduced workspace-switch stalls on desktop and Android with long chats ([#3447](https://github.com/getpaseo/paseo/pull/3447), [#3610](https://github.com/getpaseo/paseo/pull/3610))
+- Kept composer typing within the frame budget on web and desktop ([#3450](https://github.com/getpaseo/paseo/pull/3450))
+- Kept large diffs responsive while expanding, scrolling, and commenting ([#3422](https://github.com/getpaseo/paseo/pull/3422))
+- Kept large read-only source previews responsive with bounded rendering ([#3665](https://github.com/getpaseo/paseo/pull/3665))
+- Removed complete-transcript disk rewrites from long-running agent timelines ([#3647](https://github.com/getpaseo/paseo/pull/3647))
+- Added outcome summaries and failure-first grouping to pull request checks ([#3483](https://github.com/getpaseo/paseo/pull/3483))
+- Moved subagent and task trackers into pills above the composer ([#3482](https://github.com/getpaseo/paseo/pull/3482))
+- Simplified mobile agent configuration into one options sheet ([#3424](https://github.com/getpaseo/paseo/pull/3424))
+- Opened summarized tool-call groups in live sheets on compact layouts ([#3619](https://github.com/getpaseo/paseo/pull/3619))
+- Added separate Content text sizing for chat, the composer, Markdown, and review prose ([#3637](https://github.com/getpaseo/paseo/pull/3637))
+- Updated Pi usage and context meters during active turns ([#3532](https://github.com/getpaseo/paseo/pull/3532))
+- Changed Archive finished to archive every finished subagent in the track ([#3368](https://github.com/getpaseo/paseo/pull/3368))
+- Changed automatic Setup tabs to appear only after workspace setup fails ([#3634](https://github.com/getpaseo/paseo/pull/3634))
+- Changed Default send in Settings to a dropdown menu ([#3644](https://github.com/getpaseo/paseo/pull/3644))
+- Added password-equivalent warnings to pairing links in the desktop app and CLI ([#3734](https://github.com/getpaseo/paseo/pull/3734))
+- Corrected Russian UI translations and terminology ([#3586](https://github.com/getpaseo/paseo/pull/3586) by [@timz](https://github.com/timz))
+
+### Fixed
+
+- Fixed CJK IME composition being cancelled in text fields and mobile terminals ([#2811](https://github.com/getpaseo/paseo/pull/2811), [#3343](https://github.com/getpaseo/paseo/pull/3343), [#3391](https://github.com/getpaseo/paseo/pull/3391), [#3462](https://github.com/getpaseo/paseo/pull/3462), [#3517](https://github.com/getpaseo/paseo/pull/3517) by [@northsea4](https://github.com/northsea4), [@jimersylee](https://github.com/jimersylee), [@ZacharyZcR](https://github.com/ZacharyZcR), [@chulmin-dev](https://github.com/chulmin-dev))
+- Fixed OpenCode turns failing when their event stream dropped ([#3395](https://github.com/getpaseo/paseo/pull/3395))
+- Fixed OpenCode agents timing out during slow or plugin-heavy startup ([#3578](https://github.com/getpaseo/paseo/pull/3578), [#3621](https://github.com/getpaseo/paseo/pull/3621) by [@BrianAguilarWasco](https://github.com/BrianAguilarWasco))
+- Fixed an unrelated merged pull request archiving a workspace ([#3425](https://github.com/getpaseo/paseo/pull/3425))
+- Fixed Annotate element and Screenshot element doing nothing on loaded desktop browser pages ([#3187](https://github.com/getpaseo/paseo/pull/3187) by [@dgk-dev](https://github.com/dgk-dev))
+- Fixed repeated copy and fork footers after heartbeat runs ([#3484](https://github.com/getpaseo/paseo/pull/3484))
+- Fixed Cursor plan usage on hosts signed in only through `cursor-agent` ([#3486](https://github.com/getpaseo/paseo/pull/3486) by [@Lite-G](https://github.com/Lite-G))
+- Fixed mobile composers retaining sent text after submission ([#3564](https://github.com/getpaseo/paseo/pull/3564))
+- Fixed multiline composers retaining stale heights across workspace and draft lifecycles ([#3681](https://github.com/getpaseo/paseo/pull/3681), [#3740](https://github.com/getpaseo/paseo/pull/3740))
+- Fixed pull request checkout when the target repository is configured as `upstream` ([#2997](https://github.com/getpaseo/paseo/pull/2997) by [@mcowger](https://github.com/mcowger))
+- Fixed composer steers remaining unread while Claude or Codex waited on a permission ([#3585](https://github.com/getpaseo/paseo/pull/3585) by [@cleiter](https://github.com/cleiter))
+- Fixed Changes opening an empty comparison after checkout state changes ([#3636](https://github.com/getpaseo/paseo/pull/3636))
+- Fixed Pi chats ending in an error state during a successful automatic retry ([#3639](https://github.com/getpaseo/paseo/pull/3639))
+- Fixed Claude usage appearing unavailable when macOS Keychain contains multiple credential items ([#3597](https://github.com/getpaseo/paseo/pull/3597) by [@t-benoit](https://github.com/t-benoit))
+- Fixed provider usage checks overwriting Claude or Codex credential files ([#3442](https://github.com/getpaseo/paseo/pull/3442) by [@danberindei](https://github.com/danberindei))
+- Fixed active Codex agents failing to open when their native thread was archived outside Paseo ([#3334](https://github.com/getpaseo/paseo/pull/3334))
+- Fixed agents remaining impossible to stop after the provider had already settled their turn ([#3742](https://github.com/getpaseo/paseo/pull/3742))
+- Fixed worktree creation rejecting Git-valid branch names containing uppercase letters, underscores, or dots ([#3591](https://github.com/getpaseo/paseo/pull/3591))
+- Fixed project icons showing stale images in New Workspace ([#3600](https://github.com/getpaseo/paseo/pull/3600))
+- Fixed malformed SVG project icons crashing the iOS app ([#3711](https://github.com/getpaseo/paseo/pull/3711) by [@omercnet](https://github.com/omercnet))
+- Fixed pinning a project's only workspace hiding its New workspace action ([#3722](https://github.com/getpaseo/paseo/pull/3722) by [@yugui923](https://github.com/yugui923))
+- Fixed workspace layout changes bypassing timeline retry backoff ([#3736](https://github.com/getpaseo/paseo/pull/3736))
+- Fixed fresh installs resolving provider SDK versions that were not validated with the release ([#3678](https://github.com/getpaseo/paseo/pull/3678))
+
+## 0.4.0 - 2026-08-13
+
+### Breaking
+
+- Removed `paseo chat` and `paseo loop` ([#3053](https://github.com/getpaseo/paseo/pull/3053))
+
+### Added
+
+- Added host-wide agent profiles for reusable provider, model, mode, thinking, and feature settings ([#3208](https://github.com/getpaseo/paseo/pull/3208))
+- Added workspace file search to Cmd/Ctrl+P ([#3059](https://github.com/getpaseo/paseo/pull/3059))
+- Added file and folder actions to Files and Changes ([#3027](https://github.com/getpaseo/paseo/pull/3027) by [@nikuscs](https://github.com/nikuscs))
+- Added interactive Mermaid diagrams to chats and Markdown previews ([#2306](https://github.com/getpaseo/paseo/pull/2306) by [@dmeledon](https://github.com/dmeledon))
+- Added live task progress above the composer and in the timeline ([#3227](https://github.com/getpaseo/paseo/pull/3227))
+- Added a model setting for generated workspace titles, branch names, commits, and pull request drafts ([#3215](https://github.com/getpaseo/paseo/pull/3215))
+- Added the supported `@getpaseo/client` TypeScript SDK ([#3141](https://github.com/getpaseo/paseo/pull/3141))
+- Added provider installation diagnostics to the CLI ([#3243](https://github.com/getpaseo/paseo/pull/3243))
+- Added workspace rename to the CLI ([#3209](https://github.com/getpaseo/paseo/pull/3209) by [@martinhanzik](https://github.com/martinhanzik))
+
+### Improved
+
+- Added a centered reading layout and formatted YAML front matter to Markdown previews ([#3240](https://github.com/getpaseo/paseo/pull/3240))
+- Added sidebar grouping to the Command Center ([#3063](https://github.com/getpaseo/paseo/pull/3063) by [@cleiter](https://github.com/cleiter))
+- Added every built-in theme to theme shortcuts ([#3214](https://github.com/getpaseo/paseo/pull/3214))
+- Removed periodic Git polling for idle workspaces ([#3323](https://github.com/getpaseo/paseo/pull/3323))
+- Set provider catalog refreshes to a configurable two-minute deadline ([#3322](https://github.com/getpaseo/paseo/pull/3322))
+
+### Fixed
+
+- Fixed terminal sessions being lost after host sleep or daemon worker stalls ([#3235](https://github.com/getpaseo/paseo/pull/3235), [#3263](https://github.com/getpaseo/paseo/pull/3263))
+- Fixed daemon hang when archiving a workspace ([#3107](https://github.com/getpaseo/paseo/pull/3107))
+- Fixed New Workspace crash on Android when projects span multiple hosts ([#3241](https://github.com/getpaseo/paseo/pull/3241))
+- Fixed provider cache exhausting Android local storage ([#3234](https://github.com/getpaseo/paseo/pull/3234))
+- Fixed dictation recordings being lost during connection interruptions ([#3159](https://github.com/getpaseo/paseo/pull/3159))
+- Fixed new workspaces reusing an existing worktree ([#3224](https://github.com/getpaseo/paseo/pull/3224))
+- Fixed new worktrees ignoring the setup command saved in Project Settings ([#3233](https://github.com/getpaseo/paseo/pull/3233))
+- Fixed new worktrees becoming dirty when the selected base contained a different `paseo.json` ([#3311](https://github.com/getpaseo/paseo/pull/3311))
+- Fixed removed hosts continuing to receive push notifications ([#3176](https://github.com/getpaseo/paseo/pull/3176))
+- Fixed delegated-agent notifications stopping after permission prompts or workspace closure ([#3177](https://github.com/getpaseo/paseo/pull/3177), [#3192](https://github.com/getpaseo/paseo/pull/3192) by [@thomasvan](https://github.com/thomasvan), [@wilgon456](https://github.com/wilgon456))
+- Fixed OMP agent startup when model catalogs exceed the protocol-v1 frame limit ([#3184](https://github.com/getpaseo/paseo/pull/3184) by [@pi3123](https://github.com/pi3123))
+- Fixed OpenCode models being unable to return to their default variant ([#3281](https://github.com/getpaseo/paseo/pull/3281))
+- Fixed completed Codex subagents remaining marked active ([#3188](https://github.com/getpaseo/paseo/pull/3188) by [@Strainy](https://github.com/Strainy))
+- Fixed Codex compaction leaving turns marked as working ([#3211](https://github.com/getpaseo/paseo/pull/3211) by [@edihasaj](https://github.com/edihasaj))
+- Fixed Cursor usage with current Cursor logins ([#2704](https://github.com/getpaseo/paseo/pull/2704) by [@QuteSaltyFish](https://github.com/QuteSaltyFish))
+- Fixed SVG project icons rendering as blank on mobile ([#2579](https://github.com/getpaseo/paseo/pull/2579) by [@colonelpanic8](https://github.com/colonelpanic8))
+- Fixed crash when persisted cache was incompatible ([#3289](https://github.com/getpaseo/paseo/pull/3289))
+- Fixed open delegated agents being archived with their parent ([#3279](https://github.com/getpaseo/paseo/pull/3279))
+- Fixed agent creation when profiles carried modes unsupported by the selected provider ([#3331](https://github.com/getpaseo/paseo/pull/3331))
+- Fixed Claude agents waiting on Paseo's ten-minute MCP timeout ([#3315](https://github.com/getpaseo/paseo/pull/3315))
+- Fixed Copy resume command for Hermes agents ([#3300](https://github.com/getpaseo/paseo/pull/3300) by [@desmond-rai](https://github.com/desmond-rai))
+
+## 0.3.1 - 2026-08-09
+
+### Added
+
+- Unassign keyboard shortcuts and see the effective bindings in shortcut help ([#2510](https://github.com/getpaseo/paseo/pull/2510) by [@cleiter](https://github.com/cleiter), [#2985](https://github.com/getpaseo/paseo/pull/2985) by [@cleiter](https://github.com/cleiter))
+- Choose a Pure black theme ([#3012](https://github.com/getpaseo/paseo/pull/3012) by [@shrimpwtf](https://github.com/shrimpwtf))
+
+### Improved
+
+- Running agents use a clearer animated status ring ([c0daf8e](https://github.com/getpaseo/paseo/commit/c0daf8e066e0d7379723a2a6ddea20afc71ead20) by [@cleiter](https://github.com/cleiter))
+- Workspace Git updates stay responsive under file-watcher failures, heavy Git activity, and large shared-worktree repositories ([#3056](https://github.com/getpaseo/paseo/pull/3056), [#3033](https://github.com/getpaseo/paseo/pull/3033) by [@dwyanewang](https://github.com/dwyanewang), [#3078](https://github.com/getpaseo/paseo/pull/3078))
+- Kimi usage refreshes correctly and shows five-hour limits and reasoning levels ([#2743](https://github.com/getpaseo/paseo/pull/2743) by [@UnbrokenHunter](https://github.com/UnbrokenHunter))
+- Resize the desktop sidebar with touch ([#2953](https://github.com/getpaseo/paseo/pull/2953) by [@colonelpanic8](https://github.com/colonelpanic8))
+
+### Fixed
+
+- Direct LAN connections work again on macOS ([#3071](https://github.com/getpaseo/paseo/pull/3071))
+- Desktop update checks no longer report an error while platform downloads are still publishing ([#3076](https://github.com/getpaseo/paseo/pull/3076))
+- Copied lists keep their markers across selection boundaries ([#3055](https://github.com/getpaseo/paseo/pull/3055))
+- Switching hosts keeps the active project selected ([#3051](https://github.com/getpaseo/paseo/pull/3051))
+- New Workspace keeps project and isolation controls inside the composer ([#3050](https://github.com/getpaseo/paseo/pull/3050))
+- Session rename is reachable from the mobile sessions menu ([#2924](https://github.com/getpaseo/paseo/pull/2924) by [@kaspesi](https://github.com/kaspesi))
+- Desktop notification sounds play again ([#2582](https://github.com/getpaseo/paseo/pull/2582) by [@gstamp](https://github.com/gstamp))
+- Codex slash commands omit disabled skills ([#2759](https://github.com/getpaseo/paseo/pull/2759) by [@jasonhnd](https://github.com/jasonhnd))
+- Exact ignored file paths resolve correctly ([#3042](https://github.com/getpaseo/paseo/pull/3042) by [@cleiter](https://github.com/cleiter))
+- Claude question notifications summarize the requested input ([#2925](https://github.com/getpaseo/paseo/pull/2925) by [@SihyeonJeon](https://github.com/SihyeonJeon))
+- The crash screen stays readable and keeps its retry action reachable ([#3043](https://github.com/getpaseo/paseo/pull/3043))
+
 ## 0.3.0 - 2026-08-08
 
 ### Added

CLAUDE.md

@@ -47,10 +47,12 @@ At the start of non-trivial work, list `docs/` and skim anything relevant to the
 | [docs/protocol-compatibility.md](docs/protocol-compatibility.md)   | Why app/daemon versions drift, protocol vs feature contract, capability gating, COMPAT tagging                                 |
 | [docs/protocol-validation.md](docs/protocol-validation.md)         | zod-aot generated inbound WebSocket validation, patched compiler regressions, schema-purity rules                              |
 | [docs/terminal-performance.md](docs/terminal-performance.md)       | Terminal latency pipeline, coalescing/backpressure invariants, benchmark + perf spec usage                                     |
+| [docs/file-observation.md](docs/file-observation.md)               | Recursive watcher ownership, Linux constraints, teardown invariants, and Parcel comparison                                     |
 | [docs/testing.md](docs/testing.md)                                 | TDD workflow, determinism, real dependencies over mocks, test organization                                                     |
 | [docs/qa.md](docs/qa.md)                                           | QA evidence bar for pull requests — platform matrix, version drift, performance, UI proof                                      |
 | [docs/mobile-testing.md](docs/mobile-testing.md)                   | Maestro and mobile test workflows                                                                                              |
 | [docs/mobile-panels.md](docs/mobile-panels.md)                     | Compact left/center/right panel ownership, worklet motion, gesture revisions, and Fabric constraints                           |
+| [docs/explorer-sidebar.md](docs/explorer-sidebar.md)               | Explorer sidebar and ordinary side-pane host contracts, lifecycle, placement, and routing preferences                          |
 | [docs/ad-hoc-daemon-testing.md](docs/ad-hoc-daemon-testing.md)     | Isolated in-process daemon test harness                                                                                        |
 | [docs/browser-capture-harness.md](docs/browser-capture-harness.md) | Real-Electron browser screenshot harness and compositor-surface gotcha                                                         |
 | [docs/android.md](docs/android.md)                                 | App variants, local/cloud builds, EAS workflows                                                                                |

CONTRIBUTING.md

@@ -6,7 +6,7 @@ Thank you for taking the time to contribute to Paseo.
 
 Paseo is an opinionated product, built on freedom and flexibility: any agent provider, any device, running on your own machine. It is meant to be composable, so you can build the workflow you want. Read more about the product vision [here](docs/product.md).
 
-Given Paseo's scope, contributing to to it takes a lot of context that is very hard to transfer. That's why product, design, architecture, and workflow decisions are currently all made by the maintainer.
+Given Paseo's scope, contributing to it takes a lot of context that is very hard to transfer. That's why product, design, architecture, and workflow decisions are currently all made by the maintainer.
 
 I pick what to build based on whether it fits the product, how many workflows it improves, whether it keeps things composable, whether we can hold the quality bar, and whether I want to build it.
 
@@ -72,8 +72,8 @@ Here is the criteria I use to decide:
 - PRs that were explicitly approved in a discussion are preferred.
 - Unsolicited PR can be closed without a detailed review.
 - Your PR can be narrowed, refactored, or redesigned.
-- Your PR might be accpeted but not merged immediately.
-- You wil be attributed for your work in the changelog, even if I redesign your PR
+- Your PR might be accepted but not merged immediately.
+- You will be attributed for your work in the changelog, even if I redesign your PR
 
 ## QA evidence
 

README.ja.md

@@ -118,7 +118,6 @@ npx skills add getpaseo/paseo
 どのエージェントとの会話でも使用できます。
 
 - `/paseo-handoff` — エージェント間で作業を引き継ぎます。私はこれを使って Claude で計画し、Codex に実装を引き継いでいます。
-- `/paseo-loop` — 明確な受け入れ基準に沿ってエージェントをループさせます（Ralph loops とも呼ばれます）。検証役を追加することもできます。
 - `/paseo-advisor` — 単一のエージェントをアドバイザーとして起動し、作業を委任せずにセカンドオピニオンを得ます。
 - `/paseo-committee` — 対照的な2つのエージェントで委員会を構成し、一歩引いた視点で根本原因を分析して計画を作成します。
 

README.ko.md

@@ -128,7 +128,6 @@ npx skills add getpaseo/paseo
 그런 다음 어떤 에이전트 대화에서든 아래 명령을 사용할 수 있습니다.
 
 - `/paseo-handoff` — 에이전트 간에 작업을 넘깁니다. Claude로 계획을 세운 뒤 Codex에 구현을 넘길 때 이 기능을 씁니다.
-- `/paseo-loop` — 명확한 완료 조건을 기준으로 에이전트를 반복 실행합니다(Ralph loop). 검증 에이전트를 추가할 수도 있습니다.
 - `/paseo-advisor` — 작업 자체를 넘기지 않고, 에이전트 하나를 조언자로 띄워 두 번째 의견을 받습니다.
 - `/paseo-committee` — 서로 다른 관점의 에이전트 두 개로 위원회를 구성해, 한 발 물러나 근본 원인을 분석하고 계획을 세웁니다.
 

README.md

@@ -105,7 +105,7 @@ Everything you can do in the app, you can do from the terminal.
 
 ```bash
 paseo run --provider claude/opus-4.6 "implement user authentication"
-paseo run --provider codex/gpt-5.4 --worktree feature-x "implement feature X"
+paseo run --provider codex/gpt-5.5 --worktree feature-x "implement feature X"
 
 paseo ls                           # list running agents
 paseo attach abc123                # stream live output
@@ -117,6 +117,30 @@ paseo --host workstation.local:6767 run "run the full test suite"
 
 See the [full CLI reference](https://paseo.sh/docs/cli) for more.
 
+## TypeScript SDK
+
+Build issue integrations, dashboards, and orchestration services with `@getpaseo/client`:
+
+```ts
+import { createPaseoClient } from "@getpaseo/client";
+
+const client = createPaseoClient({ url: "ws://127.0.0.1:6767/ws" });
+await client.connect();
+
+const agent = await client.agents.create({
+  config: { provider: "codex/gpt-5.5" },
+  cwd: "/Users/me/dev/storefront",
+  prompt: "Review the current diff and name the riskiest change.",
+});
+
+const result = await agent.waitForFinish();
+console.log(result.lastMessage);
+
+await client.close();
+```
+
+See the [SDK quickstart](https://paseo.sh/docs/sdk/quickstart), [recipes](https://paseo.sh/docs/sdk/recipes), and [API reference](https://paseo.sh/docs/sdk/reference).
+
 ## Skills
 
 Skills teach your agent to use Paseo to orchestrate other agents.
@@ -128,7 +152,6 @@ npx skills add getpaseo/paseo
 Then use them in any agent conversation:
 
 - `/paseo-handoff` — hand off work between agents. I use this to plan with Claude and then handoff to Codex to implement.
-- `/paseo-loop` — loop an agent against clear acceptance criteria (aka Ralph loops), optionally with a verifier.
 - `/paseo-advisor` — spin up a single agent as an advisor for a second opinion, without delegating the work itself.
 - `/paseo-committee` — form a committee of two contrasting agents to step back, do root cause analysis, and produce a plan.
 

README.zh-CN.md

@@ -118,7 +118,6 @@ npx skills add getpaseo/paseo
 然后在任意 agent 对话中使用：
 
 - `/paseo-handoff` — 在 agents 之间交接工作。我会用它先和 Claude 规划，再交给 Codex 实现。
-- `/paseo-loop` — 让 agent 按明确验收标准循环工作（也叫 Ralph loops），也可以加 verifier。
 - `/paseo-advisor` — 启动单个 agent 作为 advisor，提供第二意见，但不把工作委托出去。
 - `/paseo-committee` — 组建两个风格互补的 agents，让它们后退一步做根因分析并产出计划。
 

docker/base/Dockerfile

@@ -23,6 +23,7 @@ RUN set -eux; \
     npm pack --workspace=@getpaseo/relay --pack-destination /tmp/paseo-packs; \
     npm pack --workspace=@getpaseo/protocol --pack-destination /tmp/paseo-packs; \
     npm pack --workspace=@getpaseo/client --pack-destination /tmp/paseo-packs; \
+    npm pack --workspace=@getpaseo/plugin --pack-destination /tmp/paseo-packs; \
     npm pack --workspace=@getpaseo/server --pack-destination /tmp/paseo-packs; \
     npm pack --workspace=@getpaseo/cli --pack-destination /tmp/paseo-packs
 

docs/agent-lifecycle.md

@@ -35,7 +35,16 @@ be in flight.
 
 ### Cancellation
 
-Cancellation changes lifecycle state only after the provider acknowledges the interrupt or emits a terminal turn event. If the interrupt is rejected or times out, the agent remains `running` with its active foreground turn intact. Follow-up actions such as replacement, reload, rewind, and Stop must report that failure instead of accepting work they cannot perform. Synthesizing a local cancellation without provider acknowledgment creates a split-brain session: Paseo accepts a new prompt while the provider still owns the previous foreground turn.
+Provider interruption is idempotent at the `AgentSession` boundary. It resolves when the prior
+foreground turn can no longer run, including when the provider reports that it is already idle. It
+rejects only when the provider may still own the turn. Provider adapters translate native errors
+into that contract; lifecycle callers do not interpret provider-specific errors.
+
+After an acknowledged interrupt, the manager settles the captured run even when no terminal event
+arrives or the run was still waiting for its provider turn id. The captured run token prevents an
+older cancellation from settling a newer turn. If interruption is rejected or times out, the agent
+keeps its active foreground turn and replacement, reload, rewind, and Stop report the failure.
+Accepting new work after an ambiguous interruption would create a split-brain session.
 
 ## Relationships
 
@@ -44,11 +53,26 @@ Agents can launch other agents via the agent-scoped `create_agent` MCP tool. Age
 - **Subagents** — exist as part of the creating agent's work, appear in that agent's subagent track, and are archived with it.
 - **Detached agents** — stand on their own after an explicit detach transition, do not appear in the former parent's subagent track, and are not archived with it.
 
+Parent archive detaches a subagent instead of archiving it when either condition holds:
+
+- The child belongs to another workspace.
+- The child is currently open in an agent tab.
+
+All other children archive with the parent. After the workspace layout hydrates, the client marks
+every managed subagent present in its tabs with `paseo.open-agent-tab.<client-id>=true` through the
+generic agent metadata update. This includes background and restored tabs; navigation does not own
+the marker. Closing a tab sets that client's label to `false`. Any `true` client label keeps the child
+open. Detach clears the parent and every open-tab label. The surviving child therefore becomes a
+normal root agent immediately, and closing its still-open tab archives it.
+
 Runtime ownership is resolved from explicit workspace ID and caller context, never from `cwd`. Workspace creation is a separate operation with `local | worktree` isolation; agent creation only selects an existing workspace.
 
-Users can also detach an existing subagent from the subagents track. Detach is deliberately a manual lifecycle gesture, not an agent-facing MCP tool. It removes the `paseo.parent-agent-id` label only: it does not stop, archive, move, or restart the agent. The agent keeps its current `cwd` and `workspaceId`, leaves the former parent's track, and behaves like a root agent for tab close, workspace activity, and future parent archive.
+Users can also detach an existing subagent from the subagents track. Detach is deliberately a manual lifecycle gesture, not an agent-facing MCP tool. It removes the parent and open-tab lifecycle labels: it does not stop, archive, move, or restart the agent. The agent keeps its current `cwd` and `workspaceId`, leaves the former parent's track, and behaves like a root agent for tab close, workspace activity, and future parent archive.
 
 `notifyOnFinish` defaults to `true` for agent-scoped creation and background prompt follow-ups because most delegated work needs to report back to the creating agent. Set it to `false` only for truly fire-and-forget agents or prompts.
+Permission requests are notification checkpoints, not the end of that subscription. The caller is notified again after a permission response when the child finishes, errors, or requests another permission.
+The permission notification includes the normalized request plus the child and request IDs, so the caller can inspect it and respond without fetching agent status.
+A watched child that closes before its finish event also notifies the caller so delegated work cannot disappear silently during archive or workspace teardown.
 
 ## Provider-managed child agents
 
@@ -71,7 +95,7 @@ Archiving runs through `AgentManager.archiveAgent` (`packages/server/src/server/
 2. Set `archivedAt` and normalize `lastStatus` away from `running`/`initializing`
 3. Notify subscribers
 4. Close the runtime (kills the process if still running)
-5. **Cascade-archive children** — any agent whose `paseo.parent-agent-id` label matches the archived agent gets archived too, recursively
+5. **Resolve children** — detach cross-workspace and open-tab children; cascade-archive the rest recursively
 
 Cascade is what keeps subagent fleets from outliving their orchestrator.
 
@@ -88,7 +112,9 @@ recoverable from History. Opening one pins its tab and renders the archived-agen
 timeline catch-up may load provider history with a runtime-only `history` resume purpose, which must
 leave both Paseo's `archivedAt` and the provider's native archive state unchanged. **Unarchive** remains
 the only transition back to an interactive runtime: it runs the provider's native unarchive hook
-(including Codex `thread/unarchive`) before the normal agent resume and timeline hydration flow.
+(including Codex `thread/unarchive`) before the normal agent resume and timeline hydration flow. A
+provider session can be archived outside Paseo while its Paseo agent remains active. Interactive
+resume repairs that drift through the provider's native unarchive hook; history resume does not.
 
 Provider session connection owns every process it spawns until the session is registered with
 `AgentManager`. If initialization, persisted-session resume, or initial history hydration fails,
@@ -106,7 +132,7 @@ These are two distinct concepts that used to be conflated:
 
 Closing a tab on a **root agent** still archives — the tab is the agent's home, so closing it means "I'm done with this agent." A confirm dialog protects against archiving a running agent by accident.
 
-Closing a tab on a **subagent** (any agent with `parentAgentId`) is **layout-only**. The agent stays unarchived and stays in its parent's track. The user can re-open the tab from the track at any time. This is implemented in `handleCloseAgentTab` (`packages/app/src/screens/workspace/workspace-screen.tsx`).
+Closing a tab on a **subagent** (any agent with `parentAgentId`) is **layout-only**. The app clears the current client's open-tab label before removing the tab. Another client's open tab remains protected. The agent stays unarchived and stays in its parent's track, so a later parent archive cascades to it when no client still has it open. The user can re-open the tab from the track at any time. Single and bulk tab close apply the same policy.
 
 The asymmetry is intentional: a subagent's persistent relationship lives in the parent's track. Same-workspace subagents are not auto-opened as tabs; the user opens one from that track when needed. A cross-workspace subagent is also auto-opened as a tab in its own workspace so opening that workspace does not appear empty. It remains in the parent's track until it is actually detached.
 
@@ -120,7 +146,9 @@ Running provider-native subagents contribute `running` to the workspace owned by
 
 ## The subagents track
 
-The collapsible track above the composer in an agent's pane (`packages/app/src/subagents/track.tsx`) combines two kinds of children:
+The track is a pill at the foot of an agent's pane (`packages/app/src/subagents/track.tsx`): a count you can read at a glance, and a panel behind it — a popover on wide screens, a sheet on compact ones — holding the rows. It floats over the transcript rather than sitting in a band above the composer, so the timeline scrolls underneath it; `packages/app/src/panels/agent-tracks.tsx` owns that placement, and the pill frame is shared with the task list in `packages/app/src/composer/tracks.tsx`.
+
+The rows combine two kinds of children:
 
 - **Paseo subagents** are full managed agents. Their membership rule (`packages/app/src/subagents/select.ts`) is:
 
@@ -143,6 +171,7 @@ Claude Code announces subagent lifecycle on the SDK stream (`task_started` / `ta
 - **Not every announced task belongs in the track.** Task subagents announce as `local_agent` and workflows as `local_workflow`; a backgrounded shell announces as `local_bash` with the same `tool_use_id` shape, and ambient housekeeping sets `skip_transcript`. The Claude provider normalizes a workflow to a generic provider-subagent descriptor titled `Workflow`, using Claude's summary as its description and timeline opener. Shared storage, protocol, and UI do not distinguish it from another provider subagent.
 - **A task that was never declared gets no descriptor, by any route.** Filtered tasks still emit `task_notification`s carrying a `tool_use_id`, and still emit frames carrying `parent_tool_use_id`. Attributing either produces a descriptor with no identity and a defaulted `running` status — a nameless row that never finishes. Status, presentation updates, and sidechain frames all route through the declaration table.
 - **Task ids are session-scoped, not turn-scoped.** Cancelling a turn must not clear the routing table: a backgrounded child settles after the interrupt and needs its descriptor to still exist. Cancellation instead terminalizes the declared children that were running in the foreground, and a later `task_notification` is free to correct that guess. Backgrounded children are identified by `task_updated.patch.is_backgrounded`.
+- **A resumed task can be announced again with a new `tool_use_id`.** The first Task tool id remains the canonical descriptor and later ids are routing aliases for the same session-scoped task. The resumed prompt is added to that child timeline.
 - **Effort is only reachable through hooks.** It appears nowhere on the message stream at any depth, and the level Paseo requests is not necessarily the level that runs — a model that does not support it is silently downgraded. A hook firing inside a subagent reports the active post-downgrade level next to its `agent_id`, which is the same id `task_started` calls `task_id`.
 - **Backgrounded subagents emit no frames carrying `parent_tool_use_id` at all.** Everything keyed off that field sees nothing for one; they are visible only because the task protocol announces them.
 - **On replay, `<session>/subagents/` holds every descendant, not just this session's children.** `agent-<id>.meta.json` carries `spawnDepth`: `1` is a direct child, `2+` was spawned by another subagent and its `toolUseId` names a Task call made inside its parent's session, which nothing in this transcript can resolve. Replaying those adds rows the live stream never showed, each with no Task card and no recoverable outcome, so they render as running forever. One recorded session showed 10 subagents live and would have replayed 22.
@@ -150,9 +179,9 @@ Claude Code announces subagent lifecycle on the SDK stream (`task_started` / `ta
 
 Archived Paseo subagents disappear from the track, by design. To remove one from the track without closing its tab, use the **archive button** on the row — it opens a confirm dialog and archives the subagent on confirm. Provider-owned rows have no individual Paseo lifecycle controls.
 
-The track header's **Archive finished** action hides finished provider-owned rows in the current app session. Their native sessions and timelines are untouched, and managed Paseo subagents are not archived by this bulk action. If a hidden provider child starts running again, the app brings it back to the track.
+The **Archive finished** row at the foot of the panel covers every finished row. It archives idle or errored managed Paseo subagents one at a time, and hides completed, failed, or canceled provider-owned rows in the current app session. Native sessions and timelines are untouched. Running and initializing children remain in the track. If a hidden provider child starts running again, the app brings it back to the track.
 
-To keep the agent alive but remove it from the parent's track, use **detach**. The daemon clears the parent label, emits the normal agent update, and every client reclassifies the agent from subagent to root/sibling from that updated snapshot.
+To keep the agent alive but remove it from the parent's track, use **detach**. The daemon clears the relationship lifecycle labels, emits the normal agent update, and every client reclassifies the agent from subagent to root/sibling from that updated snapshot.
 
 ## Why this shape
 
@@ -170,7 +199,7 @@ We considered universal decoupling (no tab close ever archives, archive is alway
 
 ### Subagent accumulation under long-lived parents
 
-A parent that spawns many subagents will see the track grow. Managed Paseo subagents can be archived individually. Finished provider-owned rows can be hidden together with **Archive finished**; this is app-local presentation state and resets when the app restarts.
+A parent that spawns many subagents will see the panel's list grow; the pill only counts them. Managed Paseo subagents can be archived individually or with **Archive finished**. That action hides finished provider-owned rows locally; this presentation state resets when the app restarts.
 
 ### Cross-client tab dismissal
 
@@ -186,11 +215,12 @@ $PASEO_HOME/agents/{cwd-with-dashes}/{agent-id}.json
 
 Each agent is a single JSON file. Fields relevant to this doc:
 
-| Field                             | Type          | Meaning                                                                            |
-| --------------------------------- | ------------- | ---------------------------------------------------------------------------------- |
-| `id`                              | `string`      | Stable identifier                                                                  |
-| `archivedAt`                      | `string?`     | Soft-delete timestamp (ISO 8601)                                                   |
-| `labels["paseo.parent-agent-id"]` | `string?`     | Parent agent ID, set automatically for agent-scoped creation and removed by detach |
-| `lastStatus`                      | `AgentStatus` | `initializing` / `idle` / `running` / `error` / `closed`                           |
+| Field                                        | Type          | Meaning                                                                            |
+| -------------------------------------------- | ------------- | ---------------------------------------------------------------------------------- |
+| `id`                                         | `string`      | Stable identifier                                                                  |
+| `archivedAt`                                 | `string?`     | Soft-delete timestamp (ISO 8601)                                                   |
+| `labels["paseo.parent-agent-id"]`            | `string?`     | Parent agent ID, set automatically for agent-scoped creation and removed by detach |
+| `labels["paseo.open-agent-tab.<client-id>"]` | `string?`     | `"true"` protects an open tab on that client; detach clears every matching label   |
+| `lastStatus`                                 | `AgentStatus` | `initializing` / `idle` / `running` / `error` / `closed`                           |
 
 See [`docs/data-model.md`](./data-model.md) for the full agent record.

docs/android.md

@@ -82,11 +82,11 @@ Or from `packages/app`:
 
 ```bash
 # Debug
-npx cross-env APP_VARIANT=development expo prebuild --platform android --non-interactive
+npx cross-env APP_VARIANT=development expo prebuild --platform android --clean --non-interactive
 npx cross-env APP_VARIANT=development expo run:android --variant=debug
 
 # Release
-npx cross-env APP_VARIANT=production expo prebuild --platform android --non-interactive
+npx cross-env APP_VARIANT=production expo prebuild --platform android --clean --non-interactive
 npx cross-env APP_VARIANT=production expo run:android --variant=release
 
 # Clear generated Android project

docs/architecture.md

@@ -64,20 +64,21 @@ not retain non-Git directories.
 
 **Key modules:**
 
-| Module                          | Responsibility                                                                |
-| ------------------------------- | ----------------------------------------------------------------------------- |
-| `server/bootstrap.ts`           | Daemon initialization: HTTP server, WS server, agent manager, storage, relay  |
-| `server/websocket-server.ts`    | WebSocket connection management, hello handshake, binary frame routing        |
-| `server/session.ts`             | Per-client session state, timeline subscriptions, terminal operations         |
-| `server/agent/agent-manager.ts` | Agent lifecycle state machine, timeline tracking, subscriber management       |
-| `server/agent/agent-storage.ts` | File-backed JSON persistence at `$PASEO_HOME/agents/`                         |
-| `server/agent/tools/`           | Transport-neutral catalog for workspaces, agents, permissions, and automation |
-| `server/agent/mcp-server.ts`    | Thin MCP adapter that registers the Paseo tool catalog with the MCP SDK       |
-| `server/agent/providers/`       | Provider adapters (see "Agent providers" below)                               |
-| `server/relay-transport.ts`     | Outbound relay connection with E2E encryption                                 |
-| `server/schedule/`              | Cron-based scheduled agents                                                   |
-| `server/loop-service.ts`        | Looping agent runs that retry until an exit condition                         |
-| `server/chat/`                  | Chat rooms for agent-to-agent and human-to-agent messaging                    |
+| Module                          | Responsibility                                                                 |
+| ------------------------------- | ------------------------------------------------------------------------------ |
+| `server/bootstrap.ts`           | Daemon initialization: HTTP server, WS server, agent manager, storage, relay   |
+| `server/websocket-server.ts`    | WebSocket connection management, hello handshake, binary frame routing         |
+| `server/session.ts`             | Per-client session state, timeline subscriptions, terminal operations          |
+| `server/directory-sync/`        | Daemon-global latest-state sequences for projects, workspaces, and agents      |
+| `server/workspace-labels/`      | Host-local label catalog, assignment mutations, and explicit subscriptions     |
+| `server/agent/agent-manager.ts` | Agent lifecycle state machine, timeline tracking, subscriber management        |
+| `server/agent/agent-storage.ts` | File-backed JSON persistence at `$PASEO_HOME/agents/`                          |
+| `server/agent/tools/`           | Transport-neutral catalog for workspaces, agents, permissions, and automation  |
+| `server/agent/mcp-server.ts`    | Thin MCP adapter that registers the Paseo tool catalog with the MCP SDK        |
+| `server/agent/providers/`       | Provider adapters (see "Agent providers" below)                                |
+| `server/orchestration-skills/`  | Bundled catalog, host selection, convergence, and skill-directory transactions |
+| `server/relay-transport.ts`     | Outbound relay connection with E2E encryption                                  |
+| `server/schedule/`              | Cron-based scheduled agents                                                    |
 
 ### `packages/protocol` — Wire schemas and shared protocol types
 
@@ -93,25 +94,43 @@ facade. App and CLI may import the low-level driver from
 `@getpaseo/client/internal/daemon-client` during migration, while new SDK-shaped
 code imports from `@getpaseo/client`.
 
+`PaseoApi` is the capability-only boundary over workspaces, agents, providers, and config.
+`PaseoClient` adds connection lifecycle. App plugin surfaces borrow an API over their selected
+host's client; plugin subprocesses use the same facade over a host-owned IPC transport.
+
 ### `packages/app` — Mobile + web client (Expo)
 
 Cross-platform React Native app that connects to one or more daemons.
 
 - Expo Router navigation (`/h/[serverId]/workspace/[workspaceId]`, `/h/[serverId]/agent/[agentId]`, etc.). The `workspaceId` URL segment is an opaque workspace id, not a directly meaningful filesystem path.
 - `HostRuntimeController` manages saved host connections, reconnection, and per-host runtime state
-- `runtime/replica-cache` keeps a non-authoritative per-host display replica in AsyncStorage: only the last focused agent, its workspace, and a short timeline tail. It restores before navigation becomes ready, leaves remote hydration flags false, and is atomically replaced by the normal snapshot-plus-delta synchronization path.
+- `runtime/replica-cache` keeps the complete project, workspace, and active-agent directory plus one short focused timeline tail in AsyncStorage. It restores before navigation becomes ready and leaves remote hydration flags false.
+- `runtime/directory-sync` owns directory reconciliation. On reconnect it passes the persisted per-entity cursor through `project.list`, `fetch_workspaces`, and `fetch_agents`; the daemon returns each entity's latest projection when its sequence is newer, plus tombstones.
+- `workspace-labels` owns one sequenced catalog replica per connected host, the deterministic cross-host projection that surfaces spanning hosts use (the filter page, the manager), and the per-host resolution a workspace row's chips use. Two hosts may give one name different colors, so a row resolves against its own host's catalog and a merged answer would be wrong there. Catalogs never synchronize between hosts; assignment creates a missing definition only on the target host. On the daemon, catalog and assignment rewrites share a journaled commit boundary. Startup recovery completes that commit before workspace or catalog publication.
 - `SessionContext` wraps the daemon client for the active session
 - Composer UI and submit/draft behavior live in `packages/app/src/composer/`; screens and panels should integrate it from there instead of dropping composer internals into `components/`, `hooks/`, or `screens/workspace/`
 - Timeline reducers in `timeline/session-stream-reducers.ts` handle compaction, gap detection, sequence-based deduplication
 - Timeline sync correctness is documented in [docs/timeline-sync.md](timeline-sync.md): live streams are for immediacy, `fetch_agent_timeline_request` is authoritative, and catch-up is paged but complete.
 - Voice features: dictation (STT) and voice agent (realtime)
 
-The replica cache exists only to paint stale data immediately while the host connects. It does not
-own mutations, infer deletions, or replace daemon reconciliation. Pending permission requests are
-not restored from it. AsyncStorage is not encrypted, so the cached timeline tail may contain source
-code, prompts, and tool output; encrypted-at-rest storage is a separate product/security decision.
-Its serialized payload has a 1 MiB byte budget and evicts whole host snapshots in least-recently-
-written order; a single oversized host is omitted rather than partially restored.
+The replica cache paints stale data immediately while the host connects. Directory cursors are
+reconciliation checkpoints; cached entities remain non-authoritative until the daemon answers.
+Pending permission requests are not restored from it. AsyncStorage is not encrypted, so the cached
+timeline tail may contain source code, prompts, and tool output; encrypted-at-rest storage is a
+separate product/security decision. Its serialized payload has a 32 MiB byte budget and evicts whole
+host snapshots in least-recently-written order; a single oversized host is omitted rather than
+partially restored. Browser and Electron builds store it in IndexedDB. Native builds use
+AsyncStorage, and Android reserves 64 MiB for that database.
+
+The three directory entity types have independent monotonic sequences and share one daemon
+generation. The daemon retains only the latest projection per entity and bounded tombstones, not an
+event log. A missing, expired, or previous-generation cursor receives a full snapshot. Projects are
+independent records; a project with no workspaces does not need a workspace placeholder.
+
+Workspace label definitions use a separate, explicitly subscribed sequence. The list request both
+fetches and grants live updates for that session. A current cursor receives an empty correlated
+catch-up response when nothing changed; idle sessions and unsubscribed sessions receive no label
+traffic. Workspace assignments stay on the workspace directory sequence.
 
 ### `packages/cli` — Command-line client
 
@@ -119,13 +138,12 @@ Commander.js CLI with Docker-style commands. Common agent operations are also ex
 
 - `paseo agent ls/run/import/attach/logs/stop/delete/send/inspect/wait/archive/reload/update/mode`
 - `paseo daemon start/stop/restart/status/pair/set-password`
-- `paseo chat ls/create/inspect/post/read/wait/delete`
 - `paseo terminal ls/create/capture/send-keys/kill`
 - `paseo script ls/start/stop`
-- `paseo loop run/ls/inspect/logs/stop`
 - `paseo schedule create/ls/inspect/update/pause/resume/run-once/logs/delete`
 - `paseo heartbeat create/update/delete`
-- `paseo workspace create/ls/archive`
+- `paseo project create/ls/rename/delete`
+- `paseo workspace create/ls/rename/archive`
 - `paseo permit allow/deny/ls`
 - `paseo provider ls/models`
 - hidden legacy `paseo worktree create/ls/archive` compatibility alias
@@ -162,6 +180,10 @@ Electron wrapper for macOS, Linux, and Windows.
 - Native file access for workspace integration
 - Same WebSocket client as mobile app
 
+The desktop does not manage agent skills. It retains one compatibility reader for the old
+`skill-selection.json`, imports that preference into its managed local daemon, then deletes the old
+file after the daemon confirms persistence.
+
 **Multi-window (hybrid land-on model).** `createWindow()` in `main.ts` is reusable: `⌘⇧N`/File→New Window, relaunching the app (`second-instance`), and the sidebar "Open in new window" action each open a fresh `BrowserWindow`. Every window shows the full sidebar — there is no per-window project ownership or filtering. "Land on a project" is delivered by a per-`webContents` `PendingOpenProjectStore`: each window pulls its own pending project path on mount (`paseo:get-pending-open-project`) and runs the normal open-project flow, identical to a CLI `paseo <path>` launch.
 
 > **Window-state v1 limitation:** only the _first_ window of a session restores and persists saved geometry (size/position/maximized). Windows opened via ⌘⇧N / second-instance / "Open in new window" open at the default size, OS-cascaded, and do not persist — this avoids every window stacking on the same restored bounds and fighting over the single window-state store. Lifting this needs per-window state keys.
@@ -305,7 +327,7 @@ initializing → idle ⇄ running
   client-side dedup; the default fetch page is 200 items.
 - Timeline row `timestamp` values are canonical daemon-owned timestamps. Providers may supply original replay timestamps, but clients must not guess timestamp trust or hide time UI based on local clock heuristics.
 - Events stream to connected clients in real time; correctness is backed by authoritative timeline fetches and paged-to-completion catch-up.
-- Agent state persists to `$PASEO_HOME/agents/{cwd-with-dashes}/{agent-id}.json` (timeline rows live alongside the record). That storage path is derived from `cwd`, not from workspace id.
+- Agent state persists to `$PASEO_HOME/agents/{cwd-with-dashes}/{agent-id}.json`. Timeline rows are runtime memory; provider history is the durable transcript authority and resumed agents rebuild from it. That storage path is derived from `cwd`, not from workspace id.
 
 ## Right-sidebar boundary: directory-backed vs workspace-owned
 
@@ -380,13 +402,11 @@ Providers that can accept native tool definitions should set `supportsNativePase
 
 ```
 $PASEO_HOME/
-├── agents/{cwd-with-dashes}/{agent-id}.json   # Agent record + persisted timeline rows
+├── agents/{cwd-with-dashes}/{agent-id}.json   # Agent record
 ├── projects/projects.json                      # Project registry
 ├── projects/workspaces.json                    # Workspace registry
 ├── projects/icons/                             # Custom project icon images
-├── chat/                                       # Chat rooms
 ├── schedules/                                  # Scheduled-agent definitions and runs
-├── loops/                                      # Loop runs and logs
 ├── config.json                                 # Daemon config (mutable)
 ├── daemon-keypair.json                         # Daemon identity for relay/E2EE
 ├── push-tokens.json                            # Mobile push tokens

docs/custom-providers.md

@@ -2,7 +2,7 @@
 
 Paseo supports configuring custom agent providers through `config.json` (located at `$PASEO_HOME/config.json`, typically `~/.paseo/config.json`). You can extend built-in providers with different API backends, add ACP-compatible agents, set custom binaries, disable providers, and create multiple profiles for the same underlying provider.
 
-All provider configuration lives under `agents.providers` in config.json:
+Provider definitions live under `agents.providers` in config.json:
 
 ```json
 {
@@ -17,6 +17,20 @@ All provider configuration lives under `agents.providers` in config.json:
 
 Provider IDs must be lowercase alphanumeric with hyphens (`/^[a-z][a-z0-9-]*$/`).
 
+Each provider catalog refresh waits up to 2 minutes. If a provider loads many plugins or a large
+agent catalog during startup, raise the limit in milliseconds:
+
+```json
+{
+  "agents": {
+    "catalogRefreshTimeoutMs": 180000
+  }
+}
+```
+
+The limit applies independently to every provider refresh and covers availability plus the entire
+catalog probe. `PASEO_PROVIDER_REFRESH_TIMEOUT_MS` sets it when the config field is absent.
+
 ---
 
 ## Table of Contents
@@ -245,6 +259,8 @@ requires_openai_auth = false
 
 You can create multiple entries that extend the same built-in provider. Each gets its own entry in the provider list with independent credentials, models, and environment.
 
+"Profile" here means a provider alias, and it is not an **Agent profile** — that is a named bundle of provider, model, mode, thinking option and features, stored under `daemon.agentProfiles`. See [glossary.md](glossary.md) for all four senses of the word.
+
 Example: two different Anthropic accounts as separate profiles:
 
 ```json

docs/data-model.md

@@ -54,13 +54,11 @@ $PASEO_HOME/
 │       └── {agentId}.json               # One file per agent
 ├── schedules/
 │   └── {scheduleId}.json                # One file per schedule
-├── chat/
-│   └── rooms.json                       # All rooms + messages
-├── loops/
-│   └── loops.json                       # All loop records
 ├── projects/
 │   ├── projects.json                    # Project registry
 │   ├── workspaces.json                  # Workspace registry
+│   ├── workspace-labels.json            # Shared host-local label catalog
+│   ├── workspace-labels.transaction.json # Recoverable catalog/assignment compound commit
 │   └── icons/                           # Host-local custom project icon images
 ├── runtime/
 │   └── managed-processes/
@@ -89,7 +87,7 @@ Each agent is stored as a separate JSON file, grouped by project directory.
 | `lastActivityAt`     | `string?` (ISO 8601)                     | Last activity timestamp                                                                                                                                                                                                                                                                                                                                                             |
 | `lastUserMessageAt`  | `string?` (ISO 8601)                     | Last user message timestamp                                                                                                                                                                                                                                                                                                                                                         |
 | `title`              | `string?`                                | User-visible title                                                                                                                                                                                                                                                                                                                                                                  |
-| `labels`             | `Record<string, string>`                 | Key-value labels (default `{}`). `paseo.parent-agent-id` is set automatically for agent-scoped creation and removed by detach — see [agent-lifecycle.md](./agent-lifecycle.md)                                                                                                                                                                                                      |
+| `labels`             | `Record<string, string>`                 | Key-value labels (default `{}`). Paseo uses `paseo.parent-agent-id` for parentage and client-scoped `paseo.open-agent-tab.*` labels while managed subagent tabs are open — see [agent-lifecycle.md](./agent-lifecycle.md)                                                                                                                                                           |
 | `lastStatus`         | `AgentStatus`                            | One of: `"initializing"`, `"idle"`, `"running"`, `"error"`, `"closed"`. `closed` means the record is resumable but has no live provider runtime; archive remains represented separately by `archivedAt`.                                                                                                                                                                            |
 | `lastModeId`         | `string?`                                | Last active mode ID                                                                                                                                                                                                                                                                                                                                                                 |
 | `config`             | `SerializableConfig?`                    | Agent session configuration (see below)                                                                                                                                                                                                                                                                                                                                             |
@@ -100,7 +98,7 @@ Each agent is stored as a separate JSON file, grouped by project directory.
 | `requiresAttention`  | `boolean?`                               | Whether the agent needs user attention                                                                                                                                                                                                                                                                                                                                              |
 | `attentionReason`    | `"finished" \| "error" \| "permission"?` | Why attention is needed                                                                                                                                                                                                                                                                                                                                                             |
 | `attentionTimestamp` | `string?` (ISO 8601)                     | When attention was flagged                                                                                                                                                                                                                                                                                                                                                          |
-| `internal`           | `boolean?`                               | Whether this is a system-internal agent (loop workers, etc.)                                                                                                                                                                                                                                                                                                                        |
+| `internal`           | `boolean?`                               | Whether this is a system-internal agent                                                                                                                                                                                                                                                                                                                                             |
 | `archivedAt`         | `string?` (ISO 8601)                     | Soft-delete timestamp                                                                                                                                                                                                                                                                                                                                                               |
 
 ### Nested: SerializableConfig
@@ -179,6 +177,17 @@ Terminal activity contributes to the workspace status bucket **per `workspaceId`
 
 Single file, validated with `PersistedConfigSchema`.
 
+`agents.skills.selection` is the daemon host's orchestration-skill preference. Missing means
+`{ mode: "all" }`. Installed state is not persisted; the daemon derives it from its three managed
+skill directories and keeps config plus filesystem convergence behind one serialized owner.
+
+`paseo reload` reads and validates this file once inside the daemon. That snapshot drives resolution,
+classification, application, and reload bookkeeping. `DaemonConfigStore` owns applying runtime-safe
+fields and their removal/default semantics; session handlers and the CLI only relay the structured
+result. Normal config patches persist only the requested fields, so launch overrides and resolved
+defaults never leak into the file. Startup-only fields remain compared with the daemon's launch
+snapshot so a mixed edit can apply its live subset and still name the paths that require restart.
+
 ```
 {
   version: 1,
@@ -189,6 +198,8 @@ Single file, validated with `PersistedConfigSchema`.
     mcp: { enabled: boolean, injectIntoAgents: boolean },
     git: { maxProcessesPerSecond: number, maxProcessConcurrency: number },
     appendSystemPrompt: string,    // appended to supported provider system/developer prompts
+    terminalProfiles: TerminalProfile[],  // named shell commands; omitted means DEFAULT_TERMINAL_PROFILES
+    agentProfiles: AgentProfile[],        // named agent launch bundles; omitted means none
     cors: { allowedOrigins: string[] },
     relay: { enabled: boolean, endpoint: string, publicEndpoint: string, useTls: boolean, publicUseTls: boolean }, // new homes materialize enabled: false
     auth: { password: string }    // bcrypt hash, optional
@@ -213,6 +224,9 @@ Single file, validated with `PersistedConfigSchema`.
     local: { modelsDir: string }
   },
   agents: {
+    skills?: {
+      selection?: { mode: "all" } | { mode: "custom", skills: string[] }
+    },
     // ProviderOverrideSchema; legacy entries with `command: { mode, ... }` are migrated to the
     // current shape on load via `migrateProviderSettings`. Custom provider IDs must declare
     // `extends` (one of the built-ins or `"acp"`) and `label`. See `provider-launch-config.ts`.
@@ -221,6 +235,8 @@ Single file, validated with `PersistedConfigSchema`.
       providers: [{ provider, model?, thinkingOptionId? }]
     }
   },
+  pluginsEnabled: boolean,
+  plugins: Record<pluginId, { source: "directory", path: string, enabled?: boolean }>,
   features: {
     dictation: { enabled, stt: { provider, model, language, confidenceThreshold } },
     voiceMode: { enabled, llm, stt: { provider, model, language }, turnDetection, tts: { provider, model, voice, speakerId, speed } }
@@ -235,6 +251,21 @@ Single file, validated with `PersistedConfigSchema`.
 
 All fields are optional with sensible defaults.
 
+### Profile lists
+
+`terminalProfiles` and `agentProfiles` are both whole-list fields: a config patch replaces the
+array, never merges entries, so a client sends the complete next list on every add, edit, reorder
+and remove. List order is the display order.
+
+Absent and empty mean different things for terminal profiles — omitting the key falls back to
+`DEFAULT_TERMINAL_PROFILES`, while `[]` means the user removed them all. Agent profiles have no
+defaults, so both mean none.
+
+`PersistedConfigSchema` parses strictly, so a daemon that predates a field drops it on write
+rather than storing something it cannot describe. That is why the client gates the agent profiles
+UI on `server_info.features.agentProfiles` instead of letting a save appear to succeed against an
+older daemon.
+
 ### Git process limits
 
 Git process limits are global to one daemon. The start-rate limit defaults to `64` processes per
@@ -265,9 +296,9 @@ Environment variables override `config.json`:
 | `PASEO_GIT_MAX_PROCESS_CONCURRENCY`  | `maxProcessConcurrency`  |
 | `PASEO_GIT_CONCURRENCY`              | Legacy concurrency alias |
 
-`PASEO_GIT_MAX_PROCESS_CONCURRENCY` wins when it and the legacy alias are both set. Restart the
-daemon after changing the file or environment. Run `paseo daemon restart` for a standalone daemon.
-For a desktop-managed daemon, fully quit and reopen Paseo Desktop.
+`PASEO_GIT_MAX_PROCESS_CONCURRENCY` wins when it and the legacy alias are both set. Run `paseo reload`
+after changing `config.json`. Environment changes require a daemon restart; the launch environment
+remains authoritative during reload.
 
 `agents.metadataGeneration.providers` controls the preferred structured-generation fallback order for daemon-side metadata tasks such as commit messages, PR text, branch names, and generated agent titles. Entries are tried first in the configured order, then Paseo falls through to dynamically discovered defaults and finally the current selection when available.
 
@@ -358,133 +389,7 @@ One file per schedule. ID is 8 hex characters.
 
 ---
 
-## 4. Chat
-
-**Path:** `$PASEO_HOME/chat/rooms.json`
-
-Single file containing all rooms and messages.
-
-```json
-{
-  "rooms": [ ... ],
-  "messages": [ ... ]
-}
-```
-
-### ChatRoom
-
-| Field       | Type                | Description                         |
-| ----------- | ------------------- | ----------------------------------- |
-| `id`        | `string` (UUID)     |                                     |
-| `name`      | `string`            | Unique room name (case-insensitive) |
-| `purpose`   | `string?`           | Room description                    |
-| `createdAt` | `string` (ISO 8601) |                                     |
-| `updatedAt` | `string` (ISO 8601) | Updated on each new message         |
-
-### ChatMessage
-
-| Field              | Type                | Description                         |
-| ------------------ | ------------------- | ----------------------------------- |
-| `id`               | `string` (UUID)     |                                     |
-| `roomId`           | `string`            | FK to ChatRoom.id                   |
-| `authorAgentId`    | `string`            | Agent ID of the author              |
-| `body`             | `string`            | Message text (supports `@mentions`) |
-| `replyToMessageId` | `string?`           | FK to another ChatMessage.id        |
-| `mentionAgentIds`  | `string[]`          | Extracted `@mention` agent IDs      |
-| `createdAt`        | `string` (ISO 8601) |                                     |
-
----
-
-## 5. Loop
-
-**Path:** `$PASEO_HOME/loops/loops.json`
-
-Single file containing an array of all loop records. Writes are direct (not atomic) and serialized through an in-memory queue. On daemon startup any record with `status: "running"` is recovered as `"stopped"` with an interruption log entry.
-
-| Field                   | Type                                                | Description                                |
-| ----------------------- | --------------------------------------------------- | ------------------------------------------ |
-| `id`                    | `string`                                            | 8-char UUID prefix                         |
-| `name`                  | `string?`                                           | Human-readable name                        |
-| `prompt`                | `string`                                            | Worker prompt                              |
-| `cwd`                   | `string`                                            | Working directory                          |
-| `provider`              | `string`                                            | Default provider                           |
-| `model`                 | `string?`                                           | Default model                              |
-| `modeId`                | `string?`                                           | Default mode ID                            |
-| `workerProvider`        | `string?`                                           | Override provider for workers              |
-| `workerModel`           | `string?`                                           | Override model for workers                 |
-| `verifierProvider`      | `string?`                                           | Override provider for verifiers            |
-| `verifierModel`         | `string?`                                           | Override model for verifiers               |
-| `verifierModeId`        | `string?`                                           | Override mode ID for verifiers             |
-| `verifyPrompt`          | `string?`                                           | LLM verification prompt                    |
-| `verifyChecks`          | `string[]`                                          | Shell commands to run as checks            |
-| `archive`               | `boolean`                                           | Whether to archive worker agents after use |
-| `sleepMs`               | `number`                                            | Delay between iterations (ms)              |
-| `maxIterations`         | `number?`                                           | Cap on iterations                          |
-| `maxTimeMs`             | `number?`                                           | Total time budget (ms)                     |
-| `status`                | `"running" \| "succeeded" \| "failed" \| "stopped"` |                                            |
-| `createdAt`             | `string` (ISO 8601)                                 |                                            |
-| `updatedAt`             | `string` (ISO 8601)                                 |                                            |
-| `startedAt`             | `string` (ISO 8601)                                 |                                            |
-| `completedAt`           | `string?` (ISO 8601)                                |                                            |
-| `stopRequestedAt`       | `string?` (ISO 8601)                                |                                            |
-| `iterations`            | `LoopIteration[]`                                   |                                            |
-| `logs`                  | `LoopLogEntry[]`                                    |                                            |
-| `nextLogSeq`            | `number`                                            | Monotonic log sequence counter             |
-| `activeIteration`       | `number?`                                           | Currently executing iteration index        |
-| `activeWorkerAgentId`   | `string?`                                           | Currently running worker agent             |
-| `activeVerifierAgentId` | `string?`                                           | Currently running verifier agent           |
-
-### Nested: LoopIteration
-
-| Field               | Type                                                | Description              |
-| ------------------- | --------------------------------------------------- | ------------------------ |
-| `index`             | `number`                                            | 1-based iteration index  |
-| `workerAgentId`     | `string?`                                           | Agent ID of the worker   |
-| `workerStartedAt`   | `string` (ISO 8601)                                 |                          |
-| `workerCompletedAt` | `string?` (ISO 8601)                                |                          |
-| `verifierAgentId`   | `string?`                                           | Agent ID of the verifier |
-| `status`            | `"running" \| "succeeded" \| "failed" \| "stopped"` |                          |
-| `workerOutcome`     | `"completed" \| "failed" \| "canceled"?`            |                          |
-| `failureReason`     | `string?`                                           |                          |
-| `verifyChecks`      | `LoopVerifyCheckResult[]`                           | Shell check results      |
-| `verifyPrompt`      | `LoopVerifyPromptResult?`                           | LLM verification result  |
-
-### Nested: LoopLogEntry
-
-| Field       | Type                                                 |
-| ----------- | ---------------------------------------------------- |
-| `seq`       | `number` (monotonic)                                 |
-| `timestamp` | `string` (ISO 8601)                                  |
-| `iteration` | `number?`                                            |
-| `source`    | `"loop" \| "worker" \| "verifier" \| "verify-check"` |
-| `level`     | `"info" \| "error"`                                  |
-| `text`      | `string`                                             |
-
-### Nested: LoopVerifyCheckResult
-
-| Field         | Type                |
-| ------------- | ------------------- |
-| `command`     | `string`            |
-| `exitCode`    | `number`            |
-| `passed`      | `boolean`           |
-| `stdout`      | `string`            |
-| `stderr`      | `string`            |
-| `startedAt`   | `string` (ISO 8601) |
-| `completedAt` | `string` (ISO 8601) |
-
-### Nested: LoopVerifyPromptResult
-
-| Field             | Type                |
-| ----------------- | ------------------- |
-| `passed`          | `boolean`           |
-| `reason`          | `string`            |
-| `verifierAgentId` | `string?`           |
-| `startedAt`       | `string` (ISO 8601) |
-| `completedAt`     | `string` (ISO 8601) |
-
----
-
-## 6. Project Registry
+## 4. Project Registry
 
 **Path:** `$PASEO_HOME/projects/projects.json`
 
@@ -519,7 +424,7 @@ workspace together with its owning project.
 
 ---
 
-## 7. Workspace Registry
+## 5. Workspace Registry
 
 **Path:** `$PASEO_HOME/projects/workspaces.json`
 
@@ -542,10 +447,35 @@ Array of workspace records. A workspace is a specific working directory within a
 | `updatedAt`                    | `string` (ISO 8601)                             |                                                                                                                                                                                               |
 | `archivedAt`                   | `string \| null` (ISO 8601)                     | Soft-delete; required nullable                                                                                                                                                                |
 | `autoArchivedChangeRequestUrl` | `string \| null`                                | Change request whose merged state triggered auto-archive. Restore replaces it with the current merged change request, when present, so repeated snapshots cannot archive the workspace again. |
+| `labels`                       | `string[]?`                                     | Normalized display names assigned from this host's shared label catalog. Missing means unlabelled.                                                                                            |
 | `pinnedAt`                     | `string \| null` (ISO 8601)                     | Pinned-to-top-of-sidebar timestamp; null means "not pinned"                                                                                                                                   |
 
 > **Opaque-ID invariant:** `workspaceId` is opaque identity, never a filesystem path. Filesystem and git operations take `cwd`/`workspaceDirectory` only — never the id. A compatibility-only first-materialization bootstrap still groups pre-registry agent records by path and Git remote so existing installs retain their legacy records. That grouping never runs against a live registry, and its keys are not runtime project or workspace identity.
 
+### Workspace label catalog
+
+**Path:** `$PASEO_HOME/projects/workspace-labels.json`
+
+The catalog is shared by every workspace on one host. A definition contains a display name and one
+of the ten identity colour names (`WORKSPACE_LABEL_COLORS` in
+`packages/protocol/src/workspace-labels.ts`); trimmed, collapsed, case-insensitive name identity is
+unique within that file. Definitions have no portable ID or principal owner and remain after their
+last assignment. Editing a label takes a new name, a new colour, or both in one commit, so the two
+fields cannot land half-applied. Workspaces store label names, so a rename and a delete rewrite
+workspace assignments through one serialized compound commit while a recolour is catalog-only; a
+rename onto a name the host already has is refused rather than merged. A prepared
+`workspace-labels.transaction.json` contains both before and after images. The daemon writes the
+catalog and workspace files, then atomically changes the transaction to committed; that phase
+change is the durable commit point. Recovery rolls prepared transactions back before either
+directory is served. A committed marker proves both data files were already written, so recovery
+only loads the current catalog and retries marker cleanup; it never reapplies stale workspace
+after-images over later registry mutations. A request rejected before the commit point therefore
+cannot take effect after restart. If the live daemon cannot determine or restore the durable state,
+the workspace registry freezes every write and later label mutations fail with
+`workspace_label_storage_uncertain` until daemon restart performs recovery. Reads remain available,
+but may reflect the last acknowledged cache until restart. Workspace directory and catalog updates
+publish only after the commit point; publication failure does not roll durable state back.
+
 `projectId` is still a real FK: workspace records should have a matching project record. Read-only
 history surfaces tolerate transient orphaned workspaces by omitting those rows so one bad FK cannot
 blank the whole History screen, but mutation paths should repair or remove the orphaned state rather
@@ -553,7 +483,7 @@ than treating it as valid.
 
 ---
 
-## 8. Push Token Store
+## 6. Push Token Store
 
 **Path:** `$PASEO_HOME/push-tokens.json`
 
@@ -567,7 +497,7 @@ Simple set of Expo push notification tokens. Loaded with permissive parsing (fil
 
 ---
 
-## 9. Daemon meta files
+## 7. Daemon meta files
 
 These small files are not validated as full Zod schemas but are persisted under `$PASEO_HOME` for daemon identity and runtime coordination.
 

docs/design.md

@@ -28,11 +28,13 @@ Before adding a new component, read `components/ui/`. The primitive usually exis
 
 ## 3. Hierarchy
 
-Hierarchy is conveyed through weight and color, not size. Most labels, titles, and hints across the app are `fontSize.base` or `fontSize.xs`. The distinction between a row's primary line and its secondary line is `foreground` versus `foregroundMuted`.
+Hierarchy is conveyed through weight and color, not size. Most interface text is `fontSize.base`; compact metadata and hints use `fontSize.sm`. The distinction between a row's primary line and its secondary line is `foreground` versus `foregroundMuted`.
+
+The authored interface ramp uses a 14px base. New native installs default to 15px; web and desktop default to 14px. The Appearance **Interface size** setting is the rendered `fontSize.base` value and scales the rest of the UI ramp proportionally. Primary readable content has its own `fontSize.content`, which defaults to 15px on every platform. It owns message bodies, composer input, Markdown, and PR prose. Controls, navigation, metadata, tool chrome, code, diffs, editors, and terminals stay on their interface or code tokens. **Code size** remains independent.
 
 Weight has three tiers, applied by role:
 
-- **Screen titles** — the title at the top of a screen — use `<ScreenTitle>` (`packages/app/src/components/headers/screen-title.tsx`), which renders `fontSize.base` at weight `400` on compact and `300` on desktop. Top-of-screen titles are lighter on desktop, not heavier. The workspace screen header follows the same rule (`packages/app/src/screens/workspace/workspace-screen.tsx`).
+- **Screen titles** — the title in app chrome — use `<ScreenTitle>` (`packages/app/src/components/headers/screen-title.tsx`), which renders `fontSize.base` at weight `400` on compact and `300` on desktop. The New workspace hero is the only larger product title; it uses `fontSize["2xl"]` (`packages/app/src/screens/new-workspace-screen.tsx`).
 - **Structural labels** use `fontWeight.medium`. This applies to section labels above a stack of rows (`packages/app/src/components/agent-list.tsx:519-523`, `packages/app/src/components/keyboard-shortcuts-dialog.tsx:63-67`), form field labels above an input inside a modal (`packages/app/src/components/add-host-modal.tsx:19-23`, `packages/app/src/components/pair-link-modal.tsx:24-28`), the title at the top of a modal/sheet/dialog (`packages/app/src/components/adaptive-modal-sheet.tsx:90-94`, `packages/app/src/components/ui/combobox.tsx:1607-1611`, `packages/app/src/components/welcome-screen.tsx:48-53`), action button labels in tight components such as the sidebar callout actions (`packages/app/src/components/sidebar-callout.tsx:218-221`), and inline data emphasis on dense metadata rows (`packages/app/src/components/git-diff-pane.tsx:2322-2327`, `packages/app/src/components/file-explorer-pane.tsx:1115-1122`).
 - **Content** uses `fontWeight.normal`. This applies to settings rows (`packages/app/src/styles/settings.ts`), sidebar primary list-item titles (`packages/app/src/components/sidebar-workspace-list.tsx:2680-2686`, `packages/app/src/components/agent-list.tsx:572-578`), `<Button>` text (`packages/app/src/components/ui/button.tsx:80-84`), `<StatusBadge>` text (`packages/app/src/components/ui/status-badge.tsx:56-60`), and `<SidebarCallout>` titles (`packages/app/src/components/sidebar-callout.tsx:175-180`).
 
@@ -60,11 +62,15 @@ The button is `<Button>` (`packages/app/src/components/ui/button.tsx`). It has f
 
 `ghost` is structural and non-committal — no border, no fill. Back arrows, header toggles, "Load more" footers (`packages/app/src/screens/sessions-screen.tsx:54-63`), more-affordances. Ghost is used when the affordance is part of the chrome, not a decision.
 
+Header and toolbar controls use `interactionHighlight` for hovered, pressed, open, and selected
+backgrounds. It is a translucent semantic fill so the same control works over the main surface and
+the sidebar. Apply it as `backgroundColor`; setting `opacity` on the control also fades its content.
+
 `destructive` is filled with `destructive`. It only appears inside a confirm. The button on the page is `outline`; the destructive button is the confirm button inside the dialog.
 
 Sizes: `xs` for ultra-tight inline triggers. `sm` for any button sitting in a row. `md` is the page default. `lg` is reserved for large standalone CTAs.
 
-Sizes are a shared contract across control kinds, defined once in `control-geometry.ts`: `xs` = 28px tall with `fontSize.xs` labels, `sm` = 32px with `fontSize.sm`, `md`/`lg` = 44px with `fontSize.sm`. `<SegmentedControl>` (`packages/app/src/components/ui/segmented-control.tsx`) takes the same `xs`/`sm`/`md` sizes — a segmented control next to a `<Button>` of the same size always matches in height, label size, and horizontal padding. Thin chrome such as the file toolbar uses `xs`; settings rows use `sm`. Never shrink a control's font or padding locally to fit a context — if the context needs a smaller control, the size tier is missing or the wrong one is in use.
+Sizes are a shared contract across control kinds, defined once in `control-geometry.ts`: `xs` = 28px tall with `fontSize.sm` labels, `sm` = 32px with `fontSize.base`, `md`/`lg` = 44px with `fontSize.base`. `<SegmentedControl>` (`packages/app/src/components/ui/segmented-control.tsx`) takes the same `xs`/`sm`/`md` sizes — a segmented control next to a `<Button>` of the same size always matches in height, label size, and corner radius. Its segments run one padding step tighter than a button, because the gap between segments already reads as padding. The selected segment is a `surface3` fill with `foreground` text, not an inverted one — inverting it inside thin chrome puts a white slab in the toolbar. Thin chrome such as the file toolbar uses `xs`; settings rows use `sm`. Never shrink a control's font or padding locally to fit a context — if the context needs a smaller control, the size tier is missing or the wrong one is in use.
 
 A `<Pressable>` wrapping a `<Text>` is a sixth variant. It is wrong. `<Button>` accepts `style`, `textStyle`, `leftIcon`, `disabled`, `size`, and `variant`.
 
@@ -153,6 +159,8 @@ On a narrow desktop route, app navigation yields to the rendered content topolog
 
 Electron window controls are top-corner obstructions, not a compact-layout condition. Rendered surfaces declare which top corners they physically occupy; only those corners receive clearance. Full-window overlays redeclare both corners. A focused split pane owns both corners; if focus restoration temporarily exposes the full split tree, the split boundary reserves one top strip instead of assigning a control rectangle to an arbitrarily narrow leaf. The 720px desktop breakpoint preserves the default 320px sidebar and target 400px center width when the Explorer is closed; it is product policy, not an obstruction gate.
 
+Windows and Linux controls are fixed window chrome, outside scrolling header content. A tab rail that reaches them ends at their obstruction and shows the shared overflow fade. On macOS, the Explorer toggle occupies a fixed top-right window slot so opening and closing Explorer does not move the pointer target.
+
 A new list+detail feature copies the settings shell. A new workspace-shaped feature copies the workspace shell. Inventing a third shape happens in design review, not in a PR.
 
 ---
@@ -223,15 +231,15 @@ Selected state on rows in a desktop list+detail uses `surfaceSidebarHover` as th
 
 There is exactly one token per status signal — `statusSuccess`, `statusDanger`, `statusWarning`, `statusMerged` — and every status surface uses it: PR state icons, CI check icons and pies, diff stats, file-change icons, status pills, usage bars. A surface does not get a quieter or louder variant because of where it sits. If a dense list feels loud, that is a density or weight problem; fix the density, not the color. The tokens are generated, not hand-picked — see the rule in `packages/app/src/styles/theme.ts` and regenerate rather than nudging one value. The level is set by the densest consumer, the sidebar workspace list.
 
-Status **dots** are the one exception, and they are a family of their own — `statusDotSuccess`, `statusDotDanger`, `statusDotWarning`, `statusDotRunning`, read only by `getStatusDotColor` (`packages/app/src/utils/status-dot-color.ts`). Same hues and the same generation rule, but their own band: pushed away from the surface rather than toward it (darker in light, lighter in dark) and at 90% of gamut chroma against the status family's 55–60%. A dot is a few points of solid color with no shape to read and no label attached, and the running one pulses, so at the status band's lightness the dots read dimmer than the metadata beside them — backwards, since the dot is the row's state. All four move together; regenerate the set, never one hue.
+Status **dots** are the one exception, and they are a family of their own — `statusDotSuccess`, `statusDotDanger`, `statusDotWarning`, `statusDotRunning`, read only by `getStatusDotColor` (`packages/app/src/utils/status-dot-color.ts`). Same hues and the same generation rule, but their own band: 90% of gamut chroma against the status family's 55–60%. A dot is a few points of solid color with no shape to read and no label attached, and the running one pulses, so at the status band's chroma the dots read dimmer than the metadata beside them — backwards, since the dot is the row's state. Lightness is set by hue separation rather than by distance from the surface: at 6pt four dark hues on a light surface all read as one dark blob no matter how much contrast they have. So the light band runs as bright as the contrast floor allows at L=0.62, the last step where all four clear 3:1 against the sidebar's `surface2`; the dark band sits at L=0.72, where danger turns pink above. All four move together; regenerate the set, never one hue.
 
-Status pills are the status token as foreground on a 10%-alpha tint of the same token (`${token}1a`), with a 20% border (`${token}33`). The `<StatusBadge>` primitive (`packages/app/src/components/ui/status-badge.tsx`) is canonical; a pill never reaches into `palette`.
+Status pills use the status token for text on the shared `surface3` and `border` shell. The neutral shell keeps the signal legible without manufacturing translucent colors outside the theme. The `<StatusBadge>` primitive (`packages/app/src/components/ui/status-badge.tsx`) is canonical; a pill never reaches into `palette`.
 
 Status dots — the small filled circles next to a host or agent name — are `borderRadius.full` filled with the status token. Which token a given agent state maps to is owned by `getStatusDotColor` (`packages/app/src/utils/status-dot-color.ts`); a row, a group header, and a project icon all call it rather than restating the mapping. They sit in the trailing slot of a sidebar row or as a leading marker on a status pill.
 
 Identity badges — the project icon, the sidebar host badge, and the PR-panel participant avatar — do not use the theme palette. They draw from the fixed ten-color identity table in `packages/app/src/styles/identity-colors.ts`, whose hexes are held to one contrast band so a color identifies rather than ranks. Project icons and PR avatars use it as a fill with a white letter — that is `identityColor`, one theme-independent hex per identity. Host badges use it as a _foreground_ on both the glyph and the label, which is a different contrast problem that the fill table cannot solve: no single hex clears 4.5:1 against both a near-white and a dark sidebar. Foregrounds therefore come from `identityForeground(name, colorScheme)`, one set per scheme, hue unchanged. That set is generated on the **status family's** lightness and chroma fraction, because a meta row puts a host badge beside a CI check and a diff stat, and two families at different lightness make the brighter one shout. Change the status band and this one changes with it. A host with no color assigned falls back to `foregroundMuted`. The table is theme-independent by design; do not fork it per theme, and do not add hexes to it without recomputing the band.
 
-The bespoke pills in `packages/app/src/screens/settings/host-page.tsx:97-116`, `packages/app/src/components/agent-list.tsx:607-632`, and `packages/app/src/components/sidebar-workspace-list.tsx:2889-2894` are drift to be removed. New code uses `<StatusBadge>`.
+New status pills use `<StatusBadge>`. Identity, shortcut, and interactive link badges remain separate because color does not encode status there.
 
 ---
 
@@ -242,7 +250,7 @@ The bespoke pills in `packages/app/src/screens/settings/host-page.tsx:97-116`, `
 - Bare `<Text>` for a section header inside settings. `<SettingsSection>` exists.
 - A "Settings" CTA on a detail page. Detail pages are settings; settings is reached from the sidebar, the host entry, or a row's kebab menu.
 - The word "checkout" in UI strings or identifiers. The term is "workspace".
-- New color tokens or hardcoded hex outside the palette. Status pill rgba backgrounds and the identity color table are the documented exceptions (§13), not a license.
+- New color tokens or hardcoded hex outside the palette. The identity color table is the documented exception (§13), not a license.
 - Placeholder text dimmed beyond `foregroundMuted`. No extra opacity, no italics, no ghost-text.
 - `onPointerEnter` and `onPointerLeave`. They do not fire on native iOS. Hover uses Pressable's `onHoverIn`/`onHoverOut` gated with `isHovered || isCompact || isNative`.
 - Raw DOM APIs without an `isWeb` guard.

docs/development.md

@@ -23,6 +23,10 @@ Desktop dev launches its desktop-managed daemon with `PASEO_NODE_ENV=development
 so development-only providers such as Mock Load Test are available. Packaged
 desktop launches always force the daemon to production mode.
 
+The web and desktop dev launchers pass the current Git branch to Metro as
+`EXPO_PUBLIC_PASEO_DEV_BUILD_LABEL`. The expanded desktop sidebar shows it in
+the titlebar row. Production builds leave the variable unset and show no label.
+
 `npm run dev` is only a shorthand for `npm run dev:server`. Keep `127.0.0.1:6767` for the packaged app and production-style `~/.paseo` state.
 
 ## Nix desktop package
@@ -127,9 +131,11 @@ desktop-only environment inherited by terminals opened inside Paseo from couplin
 a new worktree instance to the parent desktop instance's profile or single-instance
 lock.
 
-Electron remains in the desktop dev runner's process group. Closing or stopping the
-workspace-script terminal must terminate Electron with Metro; detaching Electron
-leaves an orphan holding the worktree's single-instance lock and broken output pipes.
+The desktop workspace script `exec`s the dev runner so the terminal owns the runner
+PID. Terminal shutdown reaches the runner as `SIGHUP`; the runner stops Metro and
+asks Electron to quit through its normal app lifecycle. Do not add an npm wrapper or
+detach Electron: either change leaves an orphan holding the worktree's single-instance
+lock and broken output pipes.
 
 With desktop dev running, verify the real BrowserWindow, titlebar clearance, fullscreen
 transition, and 751-pixel settings split with:
@@ -211,6 +217,78 @@ PASEO_PROFILE_IDLE_WAIT_MS=3000             # idle baseline before switching
 PASEO_PROFILE_DUMP_COMMITS=1                # include per-commit profiler samples
 ```
 
+For warm workspace switching, point the benchmark at an app backed by seeded
+daemon state:
+
+```bash
+PASEO_PROFILE_APP_URL=http://localhost:19010 \
+  npm run profile:workspace-switching --workspace=@getpaseo/app
+```
+
+The benchmark first warms `Cmd+1` through `Cmd+7`, then records a rapid seven-workspace
+burst. It separately warms the three-entry workspace deck LRU and records `Cmd+1` through
+`Cmd+3` without waits between keys. Both scenarios report keydown-to-activation latency and
+React commits on the same browser clock. Set `PASEO_PROFILE_WORKSPACE_DIGITS`,
+`PASEO_PROFILE_WORKSPACE_LRU_SIZE`, or `PASEO_PROFILE_WARM_QUIET_MS` to change the shape. Set
+`PASEO_PROFILE_DUMP_COMMITS=1` to include the nested component breakdown for every commit.
+Set `PASEO_PROFILE_RETAINED_REPEATS=5` to repeat the retained-LRU burst for a less noisy sample.
+Set `PASEO_PROFILE_CPU_PATH=/tmp/workspace-switch.cpuprofile` to run a separate retained-LRU
+capture after the latency scenarios and write a raw CDP CPU profile without contaminating their
+timings.
+Set `PASEO_PROFILE_TRACE_PATH=/tmp/workspace-switch.trace.json` to run another separate retained-LRU
+capture and write a Chromium Performance trace with User Timing marks and screenshots. Open the
+trace in the Chrome DevTools Performance panel. Set `PASEO_PROFILE_TRACE_INVALIDATIONS=1` only for a
+focused invalidation capture; React Native's generated stacks make that mode highly intrusive.
+Set `PASEO_PROFILE_TRACE_FOCUS=1` to include focus targets, durations, and JavaScript call stacks in
+the scenario report. This mode wraps `HTMLElement.focus`, so use it only for diagnosis.
+
+For the desktop Explorer sidebar toggle, run the app against the root checkout's daemon and use:
+
+```bash
+npm run profile:explorer-toggle --workspace=@getpaseo/app
+```
+
+The harness verifies port `6768`, opens the Paseo workspace, creates and warms the Explorer pane,
+records an idle control, then measures settled and 50 ms burst Cmd+E toggles. It reports
+input-to-DOM and input-to-paint latency, React commits, mounts, unmounts, and DOM mutations. Set
+`PASEO_PROFILE_TRACE_PATH=/tmp/explorer-toggle.trace.json` or
+`PASEO_PROFILE_CPU_PATH=/tmp/explorer-toggle.cpuprofile` for separate Chromium captures. Override
+the app URL, daemon port, workspace, or server with the corresponding `PASEO_PROFILE_*` variables.
+
+For sustained composer typing, run the paired composer-versus-textarea benchmark against a seeded
+daemon:
+
+```bash
+PASEO_PROFILE_APP_URL=http://localhost:19010 \
+  npm run profile:composer-typing --workspace=@getpaseo/app
+```
+
+The benchmark opens the first workspace, preserves its existing draft, and dispatches 300 printable
+keys at a fixed 16 ms cadence without waiting for each key to paint. It measures renderer
+`keydown` to the next paint opportunity, verifies that every character survived, alternates the
+composer and plain-textarea control across three runs, then restores the original draft. The report
+includes percentiles, frame coalescing, input processing, React work, long tasks, slow samples, and
+Playwright dispatch lateness. Set `PASEO_PROFILE_TYPING_KEYS`, `PASEO_PROFILE_TYPING_CADENCE_MS`,
+`PASEO_PROFILE_TYPING_REPEATS`, or `PASEO_PROFILE_WORKSPACE_DIGIT` to change the scenario. Optional
+`PASEO_PROFILE_CPU_PATH` and `PASEO_PROFILE_TRACE_PATH` captures run separately after the latency
+measurements so profiling overhead does not contaminate them.
+
+Set `PASEO_PROFILE_TYPING_SCENARIO=height-growth` to alternate `Shift+Enter` and printable input.
+That report includes input and composer height changes plus React work grouped into composer,
+stream, and ancestor/root scopes. Ancestor/root timings include descendant work because the Profiler
+boundaries are nested. A printable key after an empty newline should not change either height.
+
+### Preview Windows and Linux window controls on macOS
+
+Desktop development can replace native macOS traffic lights with Paseo's custom controls:
+
+```bash
+PASEO_DESKTOP_WINDOW_CONTROLS=windows npm run dev:desktop
+PASEO_DESKTOP_WINDOW_CONTROLS=linux npm run dev:desktop
+```
+
+The override is rejected in packaged builds. Restart the desktop process when changing it.
+
 ### Desktop macOS compositor watchdog
 
 macOS display sleep can leave Chromium's GPU-process display link — the vsync
@@ -260,8 +338,8 @@ the daemon-global Git process limits in `$PASEO_HOME/config.json`:
 }
 ```
 
-Restart the daemon with `paseo daemon restart`. If Paseo Desktop manages the daemon, fully quit and
-reopen the desktop app. Lower values reduce machine pressure but make Git-backed workspace state and
+Reload the daemon with `paseo reload`. Environment-variable overrides still require a restart because
+the launch environment remains authoritative. Lower values reduce machine pressure but make Git-backed workspace state and
 Git RPCs wait longer. See [Git process limits](data-model.md#git-process-limits) for defaults,
 semantics, and environment-variable overrides.
 
@@ -302,9 +380,9 @@ Lifecycle commands run in the worktree through a stable script shell: `bash`
 resolved from `PATH` on macOS/Linux, and PowerShell with `-NoProfile` on
 Windows. They inherit the daemon environment plus Paseo's lifecycle variables;
 login and interactive shell startup files are not loaded, and Bash's `BASH_ENV`
-hook is unset. Daemon-run loop verify checks and ACP single-string terminal
-commands use the same non-login Bash behavior on macOS/Linux, but preserve their
-existing `cmd.exe /c` string semantics on Windows. Service scripts are separate:
+hook is unset. ACP single-string terminal commands use the same non-login Bash
+behavior on macOS/Linux, but preserve their existing `cmd.exe /c` string semantics
+on Windows. Service scripts are separate:
 they launch in a terminal and receive the service environment described below.
 
 Because the shell differs per platform, a lifecycle command that must run
@@ -435,6 +513,16 @@ For tighter loops, you can rebuild a single workspace:
 - Changed `packages/server/src/*`, `packages/cli/src/*`, `packages/relay/src/*`, or `packages/highlight/src/*`: `npm run build:server`.
 - Changed app build dependencies: `npm run build:app-deps`.
 
+## Dependency patches
+
+`patches/*.patch` are applied by `scripts/postinstall-patches.mjs` on every install. A patch only
+runs when its package is actually present, so add the package to that script's `patchedPackages`
+list when you introduce a new patch — otherwise the file sits in `patches/` and never applies.
+Regenerate a patch with `npx patch-package <package>` after editing `node_modules/<package>`, and
+patch every build the consumers use: Metro resolves the `react-native` field of a package
+(`src/*.ts` for `react-native-svg`), while Node and Vitest resolve `main`/`module`
+(`lib/commonjs`, `lib/module`). Patching only `lib/` leaves the app bundle unfixed.
+
 ## ACP provider catalog versions
 
 The in-app ACP provider catalog pins package-runner entries (`npx`, `npm exec`,
@@ -456,7 +544,7 @@ install.
 
 Use `npm run cli` to run the in-repo CLI from source (`npx tsx packages/cli/src/index.ts`). The script wraps the CLI with `scripts/dev-home.sh`, so it automatically uses this checkout's `.dev/paseo-home` and dev daemon endpoint unless you pass an explicit override. The globally installed `paseo` binary on macOS is a symlink into the installed Paseo desktop app, not this checkout — use it to drive the desktop's built-in daemon, but use `npm run cli` when you want to talk to the CLI you are editing.
 
-Canonical automation uses `paseo workspace create/ls/archive`, `paseo heartbeat create/update/delete`, and the full `paseo schedule` group. MCP heartbeat automation is intentionally smaller: create and delete only. Detach remains an explicit user lifecycle action rather than an agent tool. `paseo run --new-workspace local|worktree` composes workspace creation with agent creation. The old `paseo worktree` and `paseo run --worktree` forms are hidden compatibility aliases.
+Canonical automation uses `paseo project create/ls/rename/delete`, `paseo workspace create/ls/rename/archive`, `paseo heartbeat create/update/delete`, and the full `paseo schedule` group. MCP heartbeat automation is intentionally smaller: create and delete only. Detach remains an explicit user lifecycle action rather than an agent tool. `paseo run --new-workspace local|worktree` composes workspace creation with agent creation. The old `paseo worktree` and `paseo run --worktree` forms are hidden compatibility aliases.
 
 ```bash
 npm run cli -- ls -a -g              # List all agents globally

docs/explorer-sidebar.md

@@ -0,0 +1,74 @@
+# Explorer sidebar and side pane
+
+The Explorer sidebar and the side pane share panel implementations, but they have different shell
+contracts.
+
+| Surface          | Purpose                      | Lifecycle                                  |
+| ---------------- | ---------------------------- | ------------------------------------------ |
+| Explorer sidebar | Files and Changes navigation | Cmd+E shows or hides the dedicated dock    |
+| Side pane        | Ordinary workspace content   | Created and closed like any workspace pane |
+
+## Panel host contract
+
+Every desktop panel registers its supported `PaneHost` values and presentation. Launchers derive
+fixed-target labels and icons from that registration, filter by host, and never substitute one
+panel type for another. Tab moves reject unsupported destinations, and placement resolves only to
+a compatible pane.
+
+Files and Changes are the Explorer defaults and its singleton navigation views. Other compatible
+tabs, including agents, terminals, files, and diffs, can move between Explorer and main panes.
+Keep panel implementations independent of either shell. `WorkspacePanelHost` owns mounting and
+retention, while each shell owns its tabs, focus, dragging, resizing, and shortcuts.
+
+## Explorer sidebar
+
+`packages/app/src/workspace-tabs/explorer-sidebar.ts` owns show, hide, toggle, and view selection.
+On desktop, the shell is rendered outside the workspace split canvas so it divides the full
+workspace, including the header. It has its own persisted width and resize handle. Main-pane splits
+never read or modify that width.
+
+The persisted layout still contains the Explorer pane so tabs survive reloads. The renderer removes
+that pane from the workspace split tree and docks it separately. Persisted identifiers retain the
+literal `"explorer"` pane id and `explorerPaneIdByWorkspace` key for compatibility.
+
+The tab rail has no inline add or close controls. Its context menu opens a New Tab launcher and
+toggles the singleton Files and Changes views. Individual tab menus close instances or move
+compatible tabs to main. Explorer tabs can be reordered, but the dock cannot be split. Selecting
+an Explorer tab does not change workspace focus.
+
+Cmd+E selects Changes for Git workspaces and Files otherwise. Compact layouts use the combined
+full-screen Explorer overlay and close it after a file opens. Wide native layouts without pane
+splits use the same combined content in a resizable inline dock; opening a file leaves that dock
+visible. Both presentations keep their selection in the panel store and reuse the layout store's
+per-workspace Explorer width. They do not create a second Explorer lifecycle.
+
+## Side pane
+
+`packages/app/src/workspace-tabs/open-beside.ts` owns content opened beside the user's work. The
+layout store remembers one ordinary pane per workspace. The first side open creates a full-height
+right split around the workspace root; later side opens reuse it.
+
+Closing the pane or moving away its final tab removes it normally and clears the remembered id. A
+later side open creates a new pane. There is no hidden side-pane lifecycle.
+
+Placement intent still controls existing tabs:
+
+| Mode      | New target                  | Existing target                   |
+| --------- | --------------------------- | --------------------------------- |
+| `pane`    | opens in the requested pane | moves to the requested pane       |
+| `prefer`  | opens in the requested pane | stays where the user placed it    |
+| `focused` | opens in the focused pane   | focuses it where it already lives |
+| `ambient` | opens in a compatible pane  | focuses it where it already lives |
+
+Explicit **Open to Side** uses `pane`. Implicit opens use `prefer`, so a preference affects only a
+new target and never yanks an existing tab out of a user-selected pane.
+
+## Routing preferences
+
+Desktop **Settings → Layout → Open in side pane** has independent switches for Explorer Files,
+Explorer Changes, chat files, diff files, subagents, pull requests, and Changes links. All switches
+default off. Mobile ignores them.
+
+Panels request an implicit open through the narrow `openPreferredTarget(target, source)` pane
+contract. Entry points outside panels use `openPreferredWorkspaceTarget`. Do not branch on a
+specific shell inside a panel.

docs/expo-router.md

@@ -17,7 +17,9 @@ Each layout owns only the routes directly inside its directory.
   `h/[serverId]/index`.
 - `packages/app/src/app/h/[serverId]/_layout.tsx` owns the host leaves with
   relative screen names: `index`, `workspace/[workspaceId]/index`,
-  `agent/[agentId]`, `sessions`, `open-project`, and `settings`.
+  `agent/[agentId]`, `sessions`, `open-project`, `settings`,
+  `plugin/[pluginId]/[surfaceId]`, and
+  `plugin/[pluginId]/[contributionKind]/[contributionId]`.
 
 Expo Router warns with `[Layout children]: No route named ...` when a layout
 registers grandchildren. Treat that warning as a route-tree bug. On native, that

docs/file-observation.md

@@ -0,0 +1,21 @@
+# File observation
+
+Use `packages/server/src/server/file-observer` for recursive filesystem observation. Create one observer service at the owning service boundary, subscribe roots through it, and close it during owner shutdown. Keep platform watchers, directory discovery, exclusion reconciliation, event batching, failure cleanup, diagnostics, and teardown inside the module. A consumer supplies a root and excluded subtrees; it must not branch on the platform or manage child watchers.
+
+Do not replace the Linux implementation with `fs.watch({ recursive: true })`. Node 22 implements Linux recursion in JavaScript by walking the tree and watching every file and directory. Paseo's directory-only Linux watcher already exhausted resources before ignored-tree pruning and a watcher cap were added in [#794](https://github.com/getpaseo/paseo/pull/794).
+
+Closing an observation must remain safe after its root is renamed or removed. Workspace archive removes owned worktrees while releasing observation references, so teardown cannot assume the watched path still exists.
+
+`unsubscribe()` is an awaited barrier. Once it resolves, no scan, reconciliation, queued event, native handle, or callback may remain. `updateIgnore()` is the equivalent barrier for excluded paths: events queued below a newly excluded root cannot be delivered after it resolves. Treat path disappearance while scanning, attaching, or receiving a watcher error as rename/archive churn. Surface permission and resource errors so WorkspaceGit can enter bounded polling.
+
+Linux topology repair is trailing-debounced and scoped to the renamed path. A file rename needs one classification stat; a new or moved directory scans only its subtree. Full-tree scans are reserved for startup and ignore-set replacement. The initial Linux scan conservatively emits existing files discovered before their directory watcher was installed; WorkspaceGit batches these into one refresh. Benchmark this startup cost with pre-populated trees.
+
+Windows and macOS use one native recursive watcher plus an ignored-pruned, directory-indexed inventory. Native events deliver immediately. Named events trigger serialized shallow scans of their parent directories; known directory events also reconcile that subtree. This recovers concrete create and delete paths omitted by coalescing without walking the whole root for each edit. Startup and exclusion changes use an awaited full inventory barrier. Activity arms a full safety audit 30 seconds after it settles, with a 30-second minimum interval and a five-minute starvation bound; pathless native events also trigger an immediate shallow root scan. Observation fails into polling rather than retaining more than 250,000 file and directory entries per root.
+
+Read aggregate health from the owning observer service. Runtime metrics include active observations, native handles, tracked native files, pending events and immediate reconciliation work, scoped and full reconciliation activity and latency, and failure counts. Delayed safety audits do not count as pending work. Closing the service releases every subscription and clears its diagnostics. Do not add path lists, watcher handles, or platform-specific controls to this interface.
+
+Git owns Git-ignore evaluation. The observer accepts absolute excluded roots and applies updates without replacing the observation or exposing its watcher topology. This keeps tracked files inside otherwise ignored directories observable and keeps Git policy out of the filesystem module.
+
+Workspace Git verifies each repository metadata subscription with a one-shot canary inside the Git directory. If the event does not round-trip through the subscription callback, treat the watcher as unavailable and enter degraded polling. Refresh working-tree Git-ignore exclusions from ignore-file events and watcher recovery, never from a healthy-watcher timer.
+
+The real-filesystem contracts and daemon auto-archive lifecycle run in the normal server test suite. Use the scripts only for manual performance and soak work: `npm run measure:file-observer --workspace=@getpaseo/server` measures burst and sustained-create behavior, and `npm run repro:file-observer-teardown --workspace=@getpaseo/server` runs the teardown soak.

docs/forms.md

@@ -61,6 +61,12 @@ complete`), keyed off the opened snapshot's serverId. Waiting for data is a
   states) is owned by `components/ui/control-geometry.ts` — controls never
   declare their own, and screens never nudge global component styles to align
   a row.
+- Every text field consumes `EditingTextInput`, directly or through a UI
+  wrapper. The editing surface owns in-progress text; React observes committed
+  edits through `onChangeText`, and programmatic mutations use the input's
+  `replaceText` command. Its props omit `value` and `defaultValue`, web IME
+  candidates stay unpublished until composition commits, and lint rejects raw
+  React Native `TextInput` imports outside the primitive.
 - The form declares one size for all fields: `sm` on desktop, `md` compact
   (`useIsCompactFormFactor`).
 - Availability hierarchy: a field whose capability doesn't apply is **hidden**

docs/glossary.md

@@ -19,19 +19,23 @@ Authoritative terminology. UI label wins. Don't invent synonyms; use what's here
 - **MR** — GitLab merge request. UI label for GitLab change requests only; do not use MR for GitHub/Gitea/Forgejo.
 - **Worktree** — Paseo-managed git worktree (`~/.paseo/worktrees/{name}`); also a `workspaceKind` value. User-facing creation treats it as the `worktree` workspace isolation choice. Code and `paseo.json` retain worktree terminology for git lifecycle implementation. Forbidden: "Checkout" as a product synonym.
 - **Repository / Remote** — Internal Git observations. They may produce mutable kind, branch, and project-grouping metadata but never the host-local project ID, root, display name, or workspace membership. No UI label.
+- **Explorer sidebar** — The dedicated navigation dock containing the Files and Changes trees. UI: "Explorer sidebar". Code: `packages/app/src/workspace-tabs/explorer-sidebar.ts`; persisted pane identifiers retain the `explorer` spelling. See [explorer-sidebar.md](explorer-sidebar.md).
+- **Side pane** — An ordinary right-side workspace pane used for content opened beside the user's work. It closes and participates in workspace splits like every other main-host pane. See [explorer-sidebar.md](explorer-sidebar.md#side-pane). Forbidden: "Side panel".
 - **Directory-backed surface** — A right-sidebar surface whose content is determined by the workspace's `cwd`, so two workspaces on the same directory see identical content: git diff/status, forge change-request info, file preview/explorer contents. Keyed by `(serverId, cwd)`, never `workspaceId`. See [architecture.md](architecture.md#right-sidebar-boundary-directory-backed-vs-workspace-owned).
 - **Workspace-owned state** — Per-workspace state that never leaks to a same-`cwd` sibling: tabs, agents, terminals, panes, title, plus review drafts, diff-mode overrides, composer attachments, and file-explorer open/expand state. Keyed by `workspaceId` (`cwd` only as a fallback for old payloads). See [architecture.md](architecture.md#right-sidebar-boundary-directory-backed-vs-workspace-owned).
 - **Workspace status bucket** — Aggregate activity signal for a workspace row. Same-`cwd` workspaces intentionally share agent and terminal status buckets, while tab, agent, and terminal visibility remains scoped by `workspaceId`.
 - **Agent session** — One running instance of an agent inside a workspace (one provider, one model, one cwd, one timeline). The conceptual unit; in the UI this opens as a tab. Moving toward this as the canonical term over "Agent". Code: `AgentSnapshotPayload` (`packages/protocol/src/messages.ts:608`).
 - **Session** — Two senses: (a) per-client connection to a daemon, internal; (b) user-facing agent session, see **Agent session**. Code: `Session` (`packages/server/src/server/session.ts`) for (a). Don't confuse with: provider-side agent session log.
-- **Profile** — Internal name for the persisted shape of a host. Code: `HostProfile` (`packages/app/src/types/host-connection.ts:37`). Never user-facing.
+- **Profile** — Internal name for the persisted shape of a host. Code: `HostProfile` (`packages/app/src/types/host-connection.ts:37`). Never user-facing. Three other things are called "profile" in this codebase, so always qualify: **Agent profile**, **Terminal profile**, and the provider aliases in [custom-providers.md](custom-providers.md). Bare "profile" is forbidden in UI copy.
+- **Agent profile** — A named launch bundle: provider, model, mode, thinking option, feature values, and free-text notes. Applying one writes those values into the **Agent controls** and is then forgotten — there is no selected profile, no default profile, and no drift tracking. The notes are written for orchestrating agents and are what the `list_profiles` MCP tool surfaces. Stored host-wide in daemon config as `agentProfiles`. UI: "Agent profiles". Code: `AgentProfile` (`packages/protocol/src/messages.ts:161`). Carries no system prompt, because a per-agent system prompt can only be set at creation. Forbidden: "Favourite", which this replaced.
+- **Terminal profile** — A named shell command a terminal can launch. Stored host-wide in daemon config as `terminalProfiles`. UI: "Terminal profiles". Code: `TerminalProfile` (`packages/protocol/src/messages.ts:135`), `resolveTerminalProfiles` (`packages/protocol/src/terminal-profiles.ts`). Also selectable as a New Workspace launch target, where the alternative is the chat agent.
 - **Provider** — Agent backend (Claude Code, Codex, Copilot, OpenCode, Pi, Oh My Pi). UI: "Provider". Code: `ProviderSnapshotEntry` (`packages/protocol/src/messages.ts:198`).
 - **Model** — A specific LLM offered by a provider. UI: "Model" / "Select model". Code: `AgentModelDefinition` (`packages/protocol/src/messages.ts:187`).
 - **Tab** — UI surface representing one session inside a workspace. Not a conceptual unit; use **Agent session** when talking about the model. Code: `WorkspaceTabDescriptor` (`packages/app/src/screens/workspace/workspace-tabs-types.ts`).
 - **Terminal** — Workspace-scoped PTY shell streamed over the binary mux channel. UI: "Terminal". Code: `TerminalStreamFrame` (`packages/protocol/src/terminal-stream-protocol.ts`).
 - **Workspace script** — A named command a workspace can run, declared in `paseo.json`. Two types: `service` and `script` (a command that runs and finishes). UI: "Scripts" for the panel and the menu, both of which list either type. Code: `WorkspaceScriptPayload` (`packages/protocol/src/messages.ts:3126`). Don't confuse with **Service**, which is one of the two types rather than a synonym for the pair.
 - **Service** — A workspace script of type `service`: long-running, may bind a port, and carries `health` of `healthy | unhealthy | null` once the daemon has observed it. The only type a sidebar workspace row reports, because a running plain script says nothing about whether the workspace is up; the row names the service and turns danger when its health fails, and where several run the failing one takes the row's single slot. UI: "Services" in the sidebar's display preferences (**Show → Services**). Code: `selectWorkspaceServiceSummary` (`packages/app/src/components/sidebar/workspace-meta-row/service-summary.ts`), `WorkspaceScriptHealthSchema` (`packages/protocol/src/messages.ts:3124`). Forbidden: "Service" as a label for the Scripts panel, which lists both types.
-- **Schedule** — Cron-style trigger that creates new agents. UI: CLI/MCP (`paseo schedule`, `create_schedule`). Don't confuse with: Heartbeat (cron prompt back into the same agent) or Loop (iterative re-execution of one agent).
+- **Schedule** — Cron-style trigger that creates new agents. UI: CLI/MCP (`paseo schedule`, `create_schedule`). Don't confuse with: Heartbeat (cron prompt back into the same agent).
 - **Heartbeat** — Ephemeral cron prompt sent back into the same agent/conversation. Agent surfaces expose create, update cron, and delete only. Use for reminders and babysitting where status should return inline.
 - **Mode** — Provider-specific operational mode (plan, default, full-access, …). UI: icon-only. Code: `modeId` in `AgentSessionConfig` (`packages/protocol/src/messages.ts:257`).
 - **Attachment** — External or local context bound to an agent prompt: forge issue/change request, review context, uploaded file, text, or image. UI: "Attach issue or PR/MR". Code: `AgentAttachment` (`packages/protocol/src/messages.ts:782`).
@@ -40,7 +44,8 @@ Authoritative terminology. UI label wins. Don't invent synonyms; use what's here
 - **Composer toolbar** — The bottom control row inside the composer input. Contains agent controls, attachment button, voice controls, and stop/send controls. Code: `leftContent`, `beforeVoiceContent`, and `rightContent` slots in `MessageInput` (`packages/app/src/composer/input/input.tsx`). Forbidden: "Status bar".
 - **Agent controls** — Provider, model, mode, thinking, and provider-feature controls for an agent or draft agent. Code: `AgentControls` / `DraftAgentControls` (`packages/app/src/composer/agent-controls/index.tsx`). Forbidden: "Agent status bar".
 - **Composer footer** — Optional area rendered below the composer input but still inside the keyboard-shifted composer layout. Code: `Composer.footer` (`packages/app/src/composer/index.tsx`).
-- **Composer track** — A contextual lane above the composer input. Specific tracks use the `<thing> track` form: **Queue track**, **Subagents track**. Code: queue track inside `Composer` (`packages/app/src/composer/index.tsx`), `SubagentsTrack` (`packages/app/src/subagents/track.tsx`).
+- **Composer track** — A contextual lane at the foot of a pane, by the composer input. Specific tracks use the `<thing> track` form: **Queue track**, **Subagents track**, **Tasks track**. The subagents and tasks tracks are pills floating over the transcript, each opening a panel with its rows; the queue track is a lane inside the composer itself. Code: queue track inside `Composer` (`packages/app/src/composer/index.tsx`), pill and panel frame in `packages/app/src/composer/tracks.tsx`, `SubagentsTrack` (`packages/app/src/subagents/track.tsx`), `AgentTaskList` (`packages/app/src/composer/task-list/index.tsx`).
+- **Steer** — A follow-up admitted into an agent's already-running turn instead of interrupting it. Interrupting the turn discards steers the provider has not read yet; there is no take-back short of that, and no provider offers one. The stop button and Escape both interrupt. A steer the provider never read keeps its transcript row, matching Codex. Sending while the agent waits on a permission denies that request first, so the message reaches the same turn instead of stalling behind it. UI: "Send and steer". Code: `steerActiveTurn` (`packages/server/src/server/agent/agent-sdk-types.ts`). Don't confuse with **Queue track**, which holds messages client-side until the turn ends and never enters the running turn. Forbidden: "Delivery", "Queued prompt".
 - **Subagent** — User-facing term for an agent session related to a parent agent session. Use **subagents** in UI copy and docs. Internal daemon/provider plumbing may say "child agent" or `child_session`, especially for provider-managed imports; do not surface "child agent" as a product term.
 - **Attachment tray** — The selected-attachments row inside the composer input, above the text input. Code: `renderAttachmentTray` (`packages/app/src/composer/index.tsx`). Forbidden: "Attachment bar".
 - **Fork** — Starting a NEW agent session seeded with a copy of another session's curated chat history. The history rides along as a `chat_history` composer attachment on a draft; the fork does **not** resume or branch the provider-side session, and the source session keeps running untouched. Two targets, one surface, one RPC (`agent.fork_context`, gated on the `agentForkContext` host feature). UI: the fork menu lives in the assistant turn footer and offers "Fork in a new tab" / "Fork in a new workspace". On a **completed** turn it sits beside the copy button and pins a boundary to that turn, so the fork stops where the turn ended. On the **in-flight** turn it sits beside the progress loader and elapsed timer, and omits the boundary entirely — `selectForkContextRows` projects the whole timeline when no boundary is given, so the fork captures everything **up to now**, including the response still streaming. That omission is what makes mid-run forking work; pinning a boundary there would silently drop the live response. Forbidden on provider-subagent (read-only) panes. Code: `useForkAgent` (`packages/app/src/hooks/use-fork-agent.ts`), `selectForkContextRows` (`packages/server/src/server/agent/activity-curator.ts`). Don't confuse with: git fork (unrelated; Paseo has no such concept), **Worktree** (filesystem isolation, not history copying).

docs/hub.md

@@ -95,11 +95,11 @@ Hub authentication rejection or close code `4403` permanently revokes the local
 daemon deletes its credential, stops reconnecting, and retains only the relationship ID, Hub origin,
 scopes, and a sanitized reason for status reporting.
 
-`paseo hub disconnect` disables socket reconnect before requesting remote revocation. If the Hub is
-offline, the daemon persists `disconnecting` and retries revocation across daemon restarts without
-opening a Hub socket. This also covers an enrollment whose request may have succeeded but whose
-response was lost. `--force` removes local authority immediately and warns that remote revocation may
-still be pending.
+`paseo hub disconnect` disables socket reconnect and execution authority before making one bounded
+remote revocation request. The daemon then removes the local relationship whether the request
+succeeds or fails. A failed request returns a warning that server-side revocation may remain pending.
+`--force` skips the remote request. Legacy persisted `disconnecting` records are removed on startup;
+the daemon does not retry revocation in the background.
 
 `paseo hub logout` removes only the active human CLI credential and preserves credentials for other origins. Interactive logout inspects and optionally disconnects a same-origin daemon before deleting the login; a failed requested disconnect preserves the login. JSON and noninteractive logout never prompt or disconnect implicitly.
 

docs/menus.md

@@ -26,6 +26,15 @@ when you have actually looked at it on a phone.
 `ContextMenu` is the exception: it defaults to `compactMode="sheet"` and enables native long press.
 Disable mobile triggering explicitly on draggable rows, where long press belongs to drag instead.
 
+## When the items differ per form factor
+
+A menu whose contents depend on what else is on screen gets one component per surface, not one
+component with `isMobile` branches inside it. The workspace header menu is the example
+(`packages/app/src/screens/workspace/workspace-header-menu.tsx`): compact has no tab strip, so it
+carries new-tab actions, while wide leaves those to the strip's `+` menu and lists workspace
+actions only. Items both surfaces share live in one component that each menu renders, from the same
+callbacks, so the two can't drift.
+
 ## Selecting an item on iOS
 
 An item that closes the menu runs its action after a fixed grace period on iOS, not immediately:
@@ -37,6 +46,12 @@ and it loses. Selections used to be dropped entirely for exactly that reason.
 The consequence for callers: an `onSelect` on iOS runs a beat after the press. Don't add a second
 delay on top of it, and don't read state that the same press mutated.
 
+This is why a row inside a menu surface goes through `selectItem` rather than calling its handler
+from its own `Pressable`. `MenuItem` is not the only row shape on the engine — `ComposerTrackRow`
+is another, with its own icons and hover-revealed actions — but both hand the press to the engine
+and take `closeOnSelect` to say whether choosing them ends the surface. A row that owns its press
+outright leaves the surface open behind whatever it opened, and skips the iOS wait.
+
 ## Pages
 
 A submenu is a page, declared as data on the surface and reached by a `MenuSubTrigger` whose `id`
@@ -62,6 +77,25 @@ Opening a submenu truncates the path to the depth of the trigger that opened it.
 sliding the pointer across a row of triggers would stack up every flyout it passed instead of
 swapping between them.
 
+## A page that takes input
+
+A menu page can hold a small form — `MenuTextField` is the field for it, drawn as a row's own
+fill so its text lands on the same rail as the labels above it. Two things have to be true for
+that page, and both are set on the page definition rather than discovered later:
+
+- **`hoverIntent: false`.** A branch you skim is opened and closed by the pointer; a form is not.
+  Without this the page opens on a pointer that was only passing through, and dismisses itself —
+  draft and all — the moment your hands move to the keyboard and the mouse drifts off the flyout.
+  While such a page is open, the whole surface stops closing on hover, because the parent flyout
+  leads back to the same dismissal.
+- **The compact presentation has to be a sheet.** `MenuTextField` resolves to
+  `BottomSheetTextInput` on compact native, which reads the sheet's context and has none in a
+  popover.
+
+The sheet keeps `keyboardBehavior="interactive"`. `extend` grows a sheet to its largest snap
+point, and with `enableDynamicSizing` that point is the content's own height — a short page does
+not grow, and the keyboard comes up over the field you are typing into.
+
 ## Hover intent
 
 A flyout **overlaps its parent by 5pt** rather than sitting beside it. With a gap there is a strip
@@ -86,6 +120,9 @@ opens on press.
 | `selected` | This is the chosen value   | A check, and nothing else |
 | `active`   | This row's submenu is open | The fill, and no check    |
 
+`selected` also announces itself as `aria-checked`, so a multi-select page is audible as the list
+of on/off things it is.
+
 A selected row does **not** get a background. A check and a fill are two separate claims about the
 same state, and showing both makes a chosen row compete with the row the pointer is actually on.
 `showSelectedCheck` moves the check to a reserved leading column when a group needs to stay

docs/plugins.md

@@ -1,80 +1,137 @@
 # Local plugins
 
-Local plugins contribute daemon RPCs, native app surfaces, and composer attachment sources from one
-`index.tsx`. Paseo executes the server contribution in a subprocess and evaluates the client
-contribution in the app runtime. Plugin code is trusted code; this first slice does not sandbox it.
+Local plugins contribute daemon RPCs, native app surfaces, workspace panels, Command Center items,
+app themes, and composer attachment sources from one `index.ts`. Paseo executes the server contribution in a
+subprocess and evaluates the client contribution in the app runtime. Plugin code is trusted code;
+this first slice does not sandbox it.
 
-## Configure a directory source
+## Install a directory source
 
-Add the plugin to the root `plugins` object in `$PASEO_HOME/config.json`:
+Create a typecheckable plugin project, install its development dependencies, then install it into
+the daemon. `init` only writes the project files; it does not run the package manager.
+
+```bash
+paseo plugin init /absolute/path/to/my-plugin
+cd /absolute/path/to/my-plugin
+npm install
+npm run typecheck
+paseo plugin install /absolute/path/to/my-plugin
+paseo plugin install /absolute/path/to/my-plugin --id another-runtime-id
+paseo plugin ls
+```
+
+The daemon stores directory sources under the root `plugins` object:
 
 ```json
 {
+  "pluginsEnabled": true,
   "plugins": {
     "my-plugin": {
       "source": "directory",
-      "path": "/absolute/path/to/my-plugin"
+      "path": "/absolute/path/to/my-plugin",
+      "enabled": true
     }
   }
 }
 ```
 
-The directory contains an identity-only manifest and one entry point:
+The plugin system is disabled unless `pluginsEnabled` is `true`. Changing that root field is
+runtime-safe: run `paseo reload` after editing `config.json`. Enabling starts every configured,
+enabled plugin; disabling tears them all down without restarting the daemon. Plugin source entries
+remain lifecycle-owned and do not reload from manual config edits.
+
+The directory contains an identity-only manifest, one entry point, and local typechecking support:
 
 ```text
 my-plugin/
   paseo-plugin.json
-  index.tsx
+  index.ts
+  main.client.tsx
+  paseo-plugin.d.ts
+  package.json
+  tsconfig.json
 ```
 
+Paseo compiles TypeScript and TSX when loading the plugin, so these packages are development dependencies only.
+The generated declaration file supplies `@getpaseo/plugin` and `@getpaseo/plugin/server` types until the
+SDK is distributed as a public package. Regenerate new plugins with the matching Paseo CLI when the
+SDK contract changes.
+
 ```json
 {
   "id": "my-plugin"
 }
 ```
 
-The manifest ID must match the config key. Restart the development daemon after changing plugin
-source or configuration; directory watching is outside the first slice.
+The config key is the runtime plugin ID. The manifest ID is the default selected during install;
+`--id` overrides it. Existing configuration is not renamed when the manifest changes, and the
+runtime does not compare the two IDs. The same directory can be installed under several config
+keys.
+
+Never enable plugins on a user's behalf without explicit permission. Before asking, check the
+target daemon's current `pluginsEnabled` value. State that plugins are trusted, unsandboxed code:
+backend code can access the daemon machine, while client contributions run inside the Paseo app.
+
+Source changes are explicit. Run `paseo plugin reload <id>` to stop and fully tear down the old
+plugin before compiling and starting from disk. A failed reload stays failed; Paseo does not restore
+the old code. Use `enable`, `disable`, and `remove` to manage one plugin. Remove deletes only its
+configuration, never its source directory. The global `pluginsEnabled` switch remains available.
+
+Server contributions can write to stdout and stderr with normal Node logging. Paseo adds `[paseo]`
+entries for loading, ready, stopping, and stopped transitions. Compilation and load failures are
+recorded as stderr entries before a subprocess exists. Inspect the recent in-memory
+tail from the host plugin settings or with `paseo plugin logs <id>`. Reload, disable, and process
+failure retain the tail; removing the plugin clears it. Daemon restarts do not retain the tail, but
+structured copies remain in `$PASEO_HOME/daemon.log`. Plugin output can contain secrets, so do not
+log credentials or tokens.
 
 ## Contribute behavior and UI
 
-Default export one contribution function. Paseo calls it with a plugin-scoped context. The compiler
-removes UI registrations from the server bundle and RPC registrations from the client bundle before
-resolving dependencies.
+Default export one contribution function from `index.ts`. Keep it to contribution wiring. Runtime
+code lives behind filename boundaries:
 
-```tsx
-import { Text } from "react-native";
-import { useQuery } from "@tanstack/react-query";
-import { z } from "zod";
-import { defineRpc, type PluginContext, useRpc } from "@paseo/plugin";
+| Suffix         | Owns                                                                 |
+| -------------- | -------------------------------------------------------------------- |
+| `*.client.tsx` | React, React Native, hooks, styles, surfaces, panels, and callbacks. |
+| `*.server.ts`  | Node APIs, filesystem and process access, credentials, and handlers. |
+| `*.shared.ts`  | Zod RPC contracts and plain values used by both runtimes.            |
 
-const greetRpc = defineRpc({
-  name: "greet",
-  input: z.object({ name: z.string() }),
-  output: z.object({ message: z.string() }),
-});
+Shared files import contracts from `@getpaseo/plugin/server`. Client files import hooks from
+`@getpaseo/plugin`. Plugin UI runs on desktop and mobile across multiple themes: color every
+`Text` from `theme.colors.foreground` or `theme.colors.foregroundMuted`, and size layout from
+`layout.compact`. See `public-docs/plugins/reference.md`.
 
-function Greeting() {
-  const greet = useRpc(greetRpc);
-  const query = useQuery({
-    queryKey: ["greeting", "Paseo"],
-    queryFn: () => greet({ name: "Paseo" }),
-  });
-  return <Text>{query.data?.message}</Text>;
-}
+| Module                    | Use it for                                               |
+| ------------------------- | -------------------------------------------------------- |
+| `@getpaseo/plugin`        | hooks and UI types                                       |
+| `@getpaseo/plugin/server` | `defineRpc`, `defineAttachmentSource`, and handler types |
+
+The compiler removes client registrations and imports from the server entry point, and server
+registrations and imports from the client entry point. Importing a `*.server` module from a client
+module, or a `*.client` module from a server module, fails compilation. Top-level React Native calls
+such as `StyleSheet.create` belong in `*.client.tsx`; placing them in `index.ts` executes them in the
+server bundle.
+
+```ts
+import type { PluginContext } from "@getpaseo/plugin";
+import { Greeting } from "./greeting.client";
+import { createGreeting } from "./greeting.server";
+import { greetRpc } from "./greeting.shared";
 
 export default function contribute(plugin: PluginContext) {
-  plugin.handle(greetRpc, async ({ name }) => ({ message: `Hello, ${name}` }));
+  plugin.handle(greetRpc, createGreeting);
   plugin.addSurface("main", Greeting);
-  plugin.addSidebarItem({
-    id: "main",
-    title: "Greeting",
-    icon: "MessageCircle",
-    surface: "main",
-  });
+  plugin.addSidebarItem({ id: "main", title: "Greeting", icon: "MessageCircle", surface: "main" });
+  return () => {};
 }
 ```
 
+The contribution function must return cleanup. Server cleanup may be async; Paseo waits for it when
+the plugin is reloaded, disabled, removed, disconnected, or shut down. Cleanup is for resources
+created by plugin code. Paseo removes registered contributions, unmounts surfaces, clears query
+state, rejects pending RPCs, closes the plugin's daemon session, and stops the subprocess. Cleanup
+errors are logged and do not interrupt host teardown.
+
 Paseo owns the route, screen header, Lucide icon validation, close action, theme DTO, layout facts,
 and render error boundary. The contributed component owns the complete body below the header.
 
@@ -82,47 +139,57 @@ RPC contracts validate inputs and outputs in both the app and plugin subprocess.
 typed async function. Use the host-provided `@tanstack/react-query` for request state and caching;
 Paseo gives each plugin installation its own query client.
 
+`usePaseo()` and the handler's `{ paseo }` context expose the same `PaseoApi`: workspaces, agents,
+providers, and daemon config. They do not expose connection lifecycle. A surface borrows the
+selected host's existing connection; switching the screen's host changes both `usePaseo()` and
+`useRpc()` to that host. An offline selected host fails there and never falls through to another
+installation. A server handler owns an IPC-backed daemon session for the life of its subprocess.
+Use plugin RPC for plugin-specific backend behavior that is not a normal Paseo operation.
+
+Each subprocess gets an exclusively owned `plugin:<id>` session. That identity is reserved from
+normal clients, never resumes another session, and is cleaned immediately on exit without reconnect
+grace. During daemon startup, plugin sessions may connect while application WebSockets remain
+paused; the daemon accepts clients only after configured plugins have settled and the initial
+catalog is complete.
+
 When the same plugin contribution exists on multiple hosts, Paseo shows it once in the sidebar and
 adds a host picker to the screen header. The selected host supplies the bundle, RPC transport, and
 query cache. Plugin code cannot address another host.
 
+Workspace panels and Command Center items remain client contributions. The daemon transports their
+compiled bundle without interpreting placement or callbacks. Panel props contain workspace and agent
+IDs. Required-selector hooks read normalized client state synchronously and use shallow equality, so a
+panel does not subscribe to fields it does not render. Command callbacks materialize their snapshots
+only when invoked. Contribution discovery and panel opening never fetch active context through plugin
+RPC. Snapshot DTOs are deeply readonly and frozen at runtime so plugin code cannot mutate normalized
+app state or a memoized selection. Panels use one persisted
+`plugin` workspace-tab target, so reload, disable, removal, and restoration resolve through the
+current installed-plugin catalog. A missing contribution renders unavailable inside the tab.
+Panels declare `locations: ["workspace", "explorer"]` to opt into Explorer hosting; omission means
+workspace only. Location controls hosting, not context. An agent panel target keeps its `agentId`
+when moved between hosts. Explorer configuration can create workspace-context panels and remove
+existing agent-context instances, but it cannot create an agent panel without an agent-aware command.
+
+Command Center callbacks use the selected host's existing `PaseoApi` for normal Paseo operations.
+They use typed plugin RPC only for plugin-specific backend work. Navigation is limited to the
+plugin's registered global surfaces and workspace panels; plugins do not receive Expo Router or
+workspace-layout store access.
+
 ## Contribute composer attachments
 
 Register a declarative attachment source backed by a plugin RPC. Paseo owns the attachment menu,
 search picker, drafts, selected pill, and submission. The plugin returns complete text snapshots;
 credentials and vendor API calls stay in the daemon handler.
 
-```tsx
-const searchIssues = defineRpc({
-  name: "issues.search",
-  input: z.object({ query: z.string() }),
-  output: z.object({
-    items: z.array(
-      z.object({
-        id: z.string(),
-        identifier: z.string(),
-        title: z.string(),
-        subtitle: z.string().optional(),
-        url: z.string().url(),
-        text: z.string(),
-        resourceType: z.string(),
-      }),
-    ),
-  }),
-});
-
-const issues = defineAttachmentSource({
-  id: "issues",
-  title: "Acme issue",
-  icon: "CircleDot",
-  pickerTitle: "Attach Acme issue",
-  searchPlaceholder: "Search by identifier or title",
-  search: searchIssues,
-});
+```ts
+import type { PluginContext } from "@getpaseo/plugin";
+import { search } from "./issues.server";
+import { issues, searchIssues } from "./issues.shared";
 
 export default function contribute(plugin: PluginContext) {
-  plugin.handle(searchIssues, ({ query }) => searchAcmeIssues(query));
+  plugin.handle(searchIssues, search);
   plugin.addAttachmentSource(issues);
+  return () => {};
 }
 ```
 
@@ -131,5 +198,28 @@ on several hosts are not coalesced. The selected snapshot submits as a text atta
 external-resource presentation, so it remains readable if the plugin is removed or an older peer
 drops the optional presentation fields.
 
-See `examples/local-plugin` for a native surface and `extensions/linear` for a complete official
-attachment-source extension.
+## Contribute a theme
+
+`addTheme` takes a small light or dark palette and a display name. Paseo expands it through the
+same semantic builders as the built-in themes, so plugins do not depend on the complete app token
+contract. Unistyles needs every theme name at `StyleSheet.configure` time, so
+`packages/app/src/styles/theme.ts` reserves one light and one dark plugin slot. The appearance
+provider rewrites the matching slot when the selection changes. See [unistyles.md](unistyles.md)
+for the runtime-patching rules the appearance settings share.
+
+`addTheme` is a client registration, so the compiler strips it from the backend bundle. A daemon
+that predates it does not, and the plugin fails to start there. Daemons advertise
+`features.pluginThemes` in `server_info`; the plugin theme catalog is the one place the app reads it, and
+a host without it contributes no themes.
+
+The selection persists as `theme: "plugin"` plus a `pluginThemeId` of `<pluginId>/theme/<themeId>`,
+so equal themes on several hosts coalesce the way sidebar contributions do. Two hosts can answer
+that id with different palettes, so picking a theme records its host through
+`rememberPluginContributionHost` and resolution prefers it; a peer connecting or dropping then does
+not repaint the app. Without a preference the sorted registry snapshot decides, so the result is
+stable rather than arrival-ordered. The app resolves that id
+against the installed catalog on every change; an id nothing contributes falls back to the default
+preference instead of painting the reserved slot's placeholder colors.
+
+See `plugin-examples/local-plugin` for a native surface, `plugin-examples/linear` for a complete
+attachment-source example, and `plugin-examples/catppuccin` for a theme.

docs/product.md

@@ -12,7 +12,7 @@ The development workflow is shifting from manually editing files to orchestratin
 
 Freedom and flexibility. Every design decision follows from this:
 
-- **Multi-provider** — Use any coding agent harness. Pick the right model for each job, switch freely as the landscape shifts. No vendor-lock in.
+- **Multi-provider** — Use any coding agent harness. Pick the right model for each job, switch freely as the landscape shifts. No vendor lock-in.
 - **Cross-device** — Desktop, mobile, web, CLI. Start work at your desk, check progress from your phone, script from the terminal.
 - **Self-hosted** — The daemon runs on your machine. Your code, your keys, your environment. No inference markup, no cloud dependency.
 - **Respectful** - No telemetry, no forced cloud, no forced accounts

docs/providers.md

@@ -81,10 +81,18 @@ OpenCode owns user message IDs. Do not pass Paseo-generated IDs to OpenCode prom
 
 `AgentManager` owns the one canonical timeline row for a foreground prompt carrying a Paseo `clientMessageId`. It records that row when `startTurn` accepts, with the wire `messageId` set to the same value. Provider adapters still emit their native user-message echo with the same `clientMessageId` when available; the manager records its provider identity on the internal row without changing or redispatching the wire item. If an adapter emits the echo before `startTurn` resolves, the manager records the provider identity with the row at acceptance. Provider adapters continue to own externally initiated user rows that have no Paseo client identity. Do not perform global transcript text dedupe.
 
+Active-turn steering is an optional `AgentSession.steerActiveTurn` operation. The manager owns admission against its exact foreground turn, canonical user-message creation, echo reconciliation, and falls back to the normal interrupt-and-replace path only when the adapter reports `unavailable`. An adapter error leaves the steer's fate ambiguous and must surface without an interrupt or retry. Codex calls `turn/steer` with the native expected turn and Paseo client user-message ID. Claude pushes an admitted steer into the exact active SDK query input; isolated control commands remain unavailable. OpenCode calls `session/prompt_async` with an OpenCode-generated message ID; the server queues the prompt while busy and the next LLM call in the same Paseo turn includes it. A missing session reports `unavailable` and uses the normal interrupt fallback.
+
+A steering adapter also owes its interrupt: stopping a turn must discard the steers the provider has not read yet, or one of them resumes the turn the user just stopped. Codex clears pending input when it aborts a turn; Claude does not, so its adapter cancels the SDK messages it queued before calling `query.interrupt()`.
+
+`SteerActiveTurnOptions.clearPendingPermissions` makes permission release part of the provider contract. A provider that accepts such a steer queues it first, denies permissions blocking its delivery, and stops once the steer is read. Steers without the flag leave permissions open. A denied plan remains in the timeline because the pending card was the only other copy of its text.
+
 Rewind accepts the canonical wire `messageId` and resolves it to the provider identity before calling the adapter. A submitted prompt cannot be rewound until its provider echo supplies that identity.
 
 Submitted user-message wire items carry the same Paseo ID in `messageId` and `clientMessageId`. Provider adapters attach `clientMessageId` only to the echo for that foreground submission; provider history and externally initiated user rows do not have a Paseo client ID.
 
+Provider adapters must terminalize every transient timeline row before emitting the turn's terminal event. Codex may omit the completed `contextCompaction` item when a turn ends during compaction, so its adapter closes any pending root compaction before forwarding `turn_completed`, `turn_failed`, or `turn_canceled`. A terminal turn must never leave the client showing an operation as still loading.
+
 Draft metadata lookups should avoid creating provider sessions when the upstream provider has top-level APIs for that metadata. Prefer `AgentClient.fetchCatalog`, `listCommands`, or `listFeatures` over creating a scratch `AgentSession`; scratch sessions can show up as empty native sessions in provider import/history UIs. `fetchCatalog` is the single discovery API for models and modes — provider implementations may use one process, separate upstream calls, or static data internally, but callers outside the provider do not get separate runtime model/mode probes. Draft command listing and scratch-session feature listing require an explicit draft model. Do not resolve a default model through catalog discovery. A client-level `listFeatures` implementation may return features from an incomplete, model-less draft and owns which features are valid in that state.
 
 Provider session import has its own contract. The picker calls `listImportableSessions` and receives rows only: provider handle, cwd, title, prompt previews, and last activity. Import calls `importSession({ providerHandleId, cwd })` for the selected row and must not call listing again. The provider returns the resumed session, storage config, persistence handle, and hydrated timeline for that one native session; `AgentManager.importProviderSession` seeds the daemon timeline and publishes the Paseo agent only after it is ready.
@@ -103,6 +111,13 @@ Daemon bootstrap reconciles that ledger in the background, without blocking star
 
 The daemon keeps provider snapshots per resolved working directory, with a separate semantic global scope for settings/provider management and requests that do not carry a cwd. Provider catalog probes receive a discriminated `FetchCatalogOptions`: `{ scope: "global", force }` for global catalog refreshes, or `{ scope: "workspace", cwd, force }` for project-scoped refreshes. Providers decide what global means for their runtime; do not infer global by comparing a cwd to the user's home directory.
 
+`ProviderSnapshotManager` owns one refresh deadline per provider. The deadline starts before the
+availability check and covers that check plus the complete catalog probe. Providers that make
+multiple catalog requests must not apply this deadline separately to each request. The manager
+aborts the shared refresh signal at the deadline. Providers name active catalog operations and
+finish subprocess, server, or session cleanup before rejecting. Timeout errors list the operations
+that were still active when the deadline expired.
+
 Snapshot reads may probe providers only while the requested cwd scope is cold. Once an entry is warm, its `ready`, `error`, or `unavailable` state stays cached until an explicit refresh. Do not add TTL revalidation, focus-triggered refreshes, selector-open refreshes, or config-reload refreshes. Selector-open refetches may read an already-loading or stale React Query, but they must not force provider probing on their own.
 
 Capable clients receive a compact, content-addressed snapshot. Model rows derive their provider from the containing entry and reference snapshot-level thinking sets. The app persists that compact shape per server and cwd, then sends its hash on the next pull; an unchanged response carries no catalog body. Keep the legacy encoding for clients without the capability. The hash covers the complete client-visible compact snapshot, including status and `fetchedAt`, so explicit refreshes invalidate it even when the discovered catalog is otherwise equal.
@@ -130,6 +145,12 @@ Keep the protocol shape provider-agnostic. Do not add provider-specific renderer
 
 Kimi Code usage follows the CLI-managed credential file at `KIMI_CODE_HOME` or `~/.kimi-code/credentials/kimi-code.json`; do not probe the legacy `~/.kimi` path as the primary source for current Kimi Code installs.
 
+Cursor usage reads the desktop `state.vscdb` token first, then `cursor-agent`'s `~/.config/cursor/auth.json`. Headless hosts only have the CLI file.
+
+### Usage fetchers are read-only on credentials
+
+A fetcher reads the provider's credential file and never writes it. On a 401 or 403 it returns `unavailable` and leaves refresh to the provider's own CLI: redeeming a refresh token in the fetcher invalidates the CLI's copy (refresh tokens are single-use), and rewriting the file through the fetcher's Zod schema drops any field the schema does not model, corrupting the file for the CLI.
+
 ---
 
 ## ACP Provider Checklist
@@ -415,6 +436,7 @@ interface AgentSession {
   readonly features?: AgentFeature[];
   run(prompt: AgentPromptInput, options?: AgentRunOptions): Promise<AgentRunResult>;
   startTurn(prompt: AgentPromptInput, options?: AgentRunOptions): Promise<{ turnId: string }>;
+  steerActiveTurn?(prompt: AgentPromptInput, options: SteerActiveTurnOptions): Promise<SteerResult>;
   subscribe(callback: (event: AgentStreamEvent) => void): () => void;
   streamHistory(): AsyncGenerator<AgentStreamEvent>;
   getRuntimeInfo(): Promise<AgentRuntimeInfo>;

docs/refactors/session-decomposition-plan.md

@@ -19,23 +19,22 @@ The first carves shipped as **deep modules with a narrow Host seam**, not the `d
 
 - **#1640 — VoiceSession** (`session/voice/voice-session.ts`, seam `VoiceSessionHost`): the STT/TTS/dictation/turn-detection subsystem. _(Originally landed at `server/voice/`; relocated under `session/` so all session subsystems share one home.)_
 - **#1644 — CheckoutSession, read side** (`session/checkout/checkout-session.ts`, seam `CheckoutSessionHost`, port `CheckoutDiffSubscriber`): status, branch validate/suggest, diff subscribe/unsubscribe, manual refresh. The workspace-git observer already delegates `emitStatusUpdate`/`scheduleDiffRefresh` to it.
+- **ScheduleSession** (`session/schedule/schedule-session.ts`, seam `ScheduleSessionHost`): scheduled-agent request handling.
 
 **Next carve — CheckoutSession mutation side (Slice 4 below).** The 17 checkout _write_ handlers still inline in `session.ts` (`dispatchCheckoutMessage`, ~2010) — branch switch/rename, commit, merge, merge-from-base, pull, push, PR create/merge, github set-auto-merge/get-check-details, PR status, PR timeline, github search, stash save/pop/list — move into the existing `session/checkout/checkout-session.ts` behind `CheckoutSessionHost`. The Slice-3 observer entanglement the table fears is already resolved: #1644 moved the status/diff read side into CheckoutSession, so the workspace observer delegates today and this no longer blocks on the WorkspaceController split.
 
 ### Why this is safe at the dispatch seam (verified)
 
-`dispatchInboundMessage` builds `a() ?? b() ?? ... ?? dispatchMiscMessage()` and short-circuits on the first non-`undefined` **Promise object** (not its resolved value). Message-type spaces are **disjoint** (no duplicate `case` labels across switches), so at most one dispatcher matches any message — collapsing to delegation cannot change which handler runs. `dispatchTerminalMessage` (2150-2153) already proves this. Two quirks preserved verbatim: schedule/\* is reached via the chat dispatcher's OWN `default` arm (2183), not the top-level `??`; and `start_workspace_script_request` (a workspace type) is special-cased before terminal delegation (2150).
+`dispatchInboundMessage` builds `a() ?? b() ?? ... ?? dispatchMiscMessage()` and short-circuits on the first non-`undefined` **Promise object** (not its resolved value). Message-type spaces are **disjoint** (no duplicate `case` labels across switches), so at most one dispatcher matches any message — collapsing to delegation cannot change which handler runs. `dispatchTerminalMessage` (2150-2153) already proves this. `start_workspace_script_request` (a workspace type) is special-cased before terminal delegation (2150).
 
 Rejected alternatives: **feature-module** (free functions + wide context bag) cannot own the live state machines (workspaceUpdatesSubscription, agentUpdatesSubscription, ~25 voice fields) and adds a competing idiom; **mixin-composition** preserves the shared-`this` god object verbatim and requires widening ~325 private fields to protected.
 
 ## Slice ordering (least-coupled first)
 
-The task recommended **git/checkout as the first slice — OVERRIDDEN.** Verification: `emitCheckoutStatusUpdate` is called from exactly ONE site (session.ts:4915), inside the workspace-owned `syncWorkspaceGitObserver` callback that ALSO fires workspace effects over shared watch-target maps. Extracting checkout first forces splitting the hardest workspace/git seam before workspace is touched. The strictly safer first cuts are **chat-schedule-loop** (only knot: `handleChatPostRequest`; touches no shared observer/git/voice state) and **provider-catalog** (one shared collaborator + injected predicates).
+The task recommended **git/checkout as the first slice — OVERRIDDEN.** Verification: `emitCheckoutStatusUpdate` is called from exactly ONE site (session.ts:4915), inside the workspace-owned `syncWorkspaceGitObserver` callback that ALSO fires workspace effects over shared watch-target maps. Extracting checkout first forces splitting the hardest workspace/git seam before workspace is touched. The safer first cut is **provider-catalog** (one shared collaborator + injected predicates).
 
 | #   | Slice                                                                         | Effort | Risk   |
 | --- | ----------------------------------------------------------------------------- | ------ | ------ |
-| 0   | Test net + disjointness tripwire (no extraction)                              | M      | low    |
-| 1   | ChatScheduleLoopController — **STOP FOR REVIEW after green**                  | M      | low    |
 | 2   | ProviderCatalogController                                                     | M      | medium |
 | 3   | Split shared workspace-git observer + agent-subscribe fan-out (no controller) | M      | high   |
 | 4   | GitCheckoutController                                                         | L      | medium |
@@ -47,34 +46,6 @@ The task recommended **git/checkout as the first slice — OVERRIDDEN.** Verific
 
 ---
 
-## Slice 0 — Test net + disjointness tripwire (prerequisite)
-
-No production code moves. Add `session.dispatch-seam.test.ts`. This is the gate the whole plan rests on, because chat/schedule/loop have **zero** handleMessage coverage today (verified).
-
-Write RED-then-GREEN against the **current in-place** Session:
-
-- `chat/post` happy path (asserts `chat/post` response emitted) + fanout-limit error path (asserts the `chat/post` error envelope, NOT a bubbled `rpc_error`).
-- one `schedule/*` and one `loop/*` round-trip.
-- a handler that throws **synchronously** emits `rpc_error{code:"handler_error"}` + an `activity_log` error frame.
-- a handler that **rejects async** emits the SAME pair.
-- a table-driven assertion that the union of all controllers' owned-type `ReadonlySet`s is pairwise disjoint and covers the dispatched `SessionInboundMessage` union (grows as controllers land).
-
-**Tests:** `session.dispatch-seam.test.ts`, `session.test.ts`.
-
----
-
-## Slice 1 — ChatScheduleLoopController ← STOP FOR HUMAN REVIEW after this ships green
-
-**Move:** all 21 handlers (`handleChat*` ×7, `handleSchedule*` ×9, `handleLoop*` ×5), the three rpc-error emitters (`emitChatRpcError`/`emitScheduleRpcError`/`emitLoopRpcError` — **kept separate, not merged**), `toScheduleSummary` → `packages/server/src/server/chat/chat-schedule-loop-controller.ts`. Collapse `dispatchChatScheduleLoopMessage` + `dispatchScheduleMessage` to `return this.chatScheduleLoopController.dispatch(msg)`.
-
-**SessionContext surface:** `emit`, `sessionLogger`, `clientId` (authorAgentId fallback), `chatService`, `scheduleService`, `loopService`, and a narrow agent-control port `{ listAgents, resolveAgentIdentifier, agentStorage.list }` for `handleChatPostRequest` mention fanout.
-
-**Owned-type set MUST include all 7 `chat/*` + 5 `loop/*` + 9 `schedule/*` types** — schedule/\* is currently routed via the chat dispatcher's own `default` arm, so it must stay inside this one controller, or schedule requests silently no-op.
-
-**Behavior note:** least-coupled domain. Move the three rpc-error emitters verbatim (they differ in default code + the `ChatServiceError` branch). **Tests:** `session.dispatch-seam.test.ts`, `loop-service.test.ts`, `session.test.ts`.
-
----
-
 ## Slice 2 — ProviderCatalogController
 
 **Move:** 7 provider handlers + `emitProviderDisabledResponse` + `getProviderSnapshotEntryForRead` → `packages/server/src/server/provider/provider-catalog-controller.ts`. Move the `providers_snapshot_update` PUSH wiring (1235-1254) into the controller's `start()`/`dispose()`. Collapse `dispatchProviderMessage`.

docs/release.md

@@ -120,7 +120,7 @@ republish the packages:
 
 ```bash
 PASEO_VERSION=$(node -p "require('./package.json').version")
-for package in highlight relay protocol client server cli; do
+for package in highlight relay protocol client plugin server cli; do
   npm dist-tag add "@getpaseo/$package@$PASEO_VERSION" beta
 done
 ```
@@ -464,26 +464,37 @@ No prefix (`v`), no extra text. `Release Notes Sync` matches the `## X.Y.Z` (or
 ## Changelog ownership
 
 - **The agent running the release writes the changelog entry — beta or stable.** The release context and final wording stay with that agent.
+- **Commit history is only an index of the changes. Never draft the changelog from commit subjects or diffs alone.** For every PR in the release range, read the full PR description and every issue it links to before deciding what changed, why users care, or how changes should be grouped. Use the implementation only to verify the resulting understanding.
 - For the first beta or a direct stable release, draft from the previous stable tag to the release source. For later betas, draft from the previous beta tag to the release source. Promotion replaces the beta series with one entry drafted from the previous stable tag to the release source. Review the result against the changelog policy below, show it to the user, and wait for approval before committing it.
 
-## Changelog voice
+## Changelog wording
 
-The changelog is shown on the Paseo homepage. Write it for **end users**, not developers.
+The changelog is shown on the Paseo homepage. Each bullet is a compact factual record of
+product behavior that changed.
 
-- **Frame everything from the user's perspective.** Describe what changed in the app, not what changed in the code. Users care that "workspaces load instantly" — not that a component no longer remounts.
-- **Never mention component names, internal modules, or implementation details.** No `WorkingIndicator`, no `accumulatedUsage`, no `reconcileAndEmitWorkspaceUpdates`. Also no "virtualized lists", no "remount", no "memoization", no "debounced", no "fuzzy ranking", no "controlled input", no "uncontrolled input" — these are implementation words masquerading as user-facing copy.
-- **Concrete WRONG → RIGHT examples** (real mistakes from past releases):
+- **Name the exact change.** Prefer `Added <capability>`, `Removed <behavior>`,
+  `Changed <behavior>`, or `Fixed <failure> when <condition>`.
+- **Keep the scope exact.** A conditional bug is not a general reliability problem. Do not
+  broaden one failure into claims that Paseo is now faster, smoother, responsive, or reliable.
+- **Use concrete product and runtime terms.** Git polling, persisted cache, provider catalog,
+  and WebSocket reconnects can identify the affected behavior. Component names, internal
+  modules, code symbols, and implementation techniques cannot: omit `WorkingIndicator`,
+  `reconcileAndEmitWorkspaceUpdates`, remounts, memoization, and controlled inputs.
+- **State the consequence only when the change itself is unclear.** Keep the condition that
+  makes the consequence true. Do not replace a precise change with a broad benefit claim.
+- **Do not invent context.** Mention an upgrade, platform, workload, or user action only when
+  the PR or linked issue establishes that scope.
 
-  | Wrong (implementation-facing)                                                       | Right (user-facing)                                         |
-  | ----------------------------------------------------------------------------------- | ----------------------------------------------------------- |
-  | Switching layouts no longer remounts the active agent                               | Splitting a pane no longer loses your scroll position       |
-  | Model, mode, and thinking pickers — searchable virtualized lists with fuzzy ranking | Mobile model selector is faster and more straightforward    |
-  | Text inputs in mobile sheets no longer flicker while typing fast                    | Typing in mobile sheets no longer flickers                  |
-  | Compact web sheets no longer crash when swiped to dismiss                           | Sheets on mobile web no longer crash when swiped to dismiss |
-  | Reduced re-renders in the agent list                                                | Agent list scrolls smoothly                                 |
-  | Added debouncing to the search input                                                | Search results no longer lag behind typing                  |
+| Avoid                                                        | Write                                                 |
+| ------------------------------------------------------------ | ----------------------------------------------------- |
+| Paseo stays responsive with many idle Git workspaces         | Removed periodic Git polling for idle workspaces      |
+| Incompatible saved app data no longer crashes after upgrades | Fixed crash when persisted cache was incompatible     |
+| Splitting layouts no longer remounts the active agent        | Fixed scroll position resetting when splitting a pane |
+| Mobile model selector is faster and more straightforward     | Added search to the mobile model selector             |
 
-  Test: would a non-developer reader recognise what changed when using the app? If they'd need an engineer to translate ("what's a remount?"), the bullet is still implementation-facing — rewrite it as the symptom the user experiences.
+Test each bullet against the source PR and issue: can a reviewer point to the exact behavior
+that changed, the failure that was fixed, or the capability that was added? If the bullet only
+claims a general improvement, rewrite it with the concrete change.
 
 - **Use the entry's release scope.** Include changes within the matching range in **Changelog scope**.
 - **Collapse internal iterations within that scope.** Present a feature added and fixed in one range as working. A later beta can describe a fix to behavior delivered in an earlier beta; promotion folds the complete beta series into the final stable behavior.
@@ -498,8 +509,7 @@ Every bullet must be scannable at a glance. The changelog is not release documen
 - **One line per bullet.** If a bullet wraps to three lines in a narrow column, it's too long.
 - **Split bullets that pack multiple distinct changes.** If a bullet uses "and", "plus", a comma list, or an em-dash to chain several independent improvements, break them into separate bullets — even when they share a theme or author. One bullet = one user-facing change.
 - **Trim qualifying clauses.** Drop "with a hint shown when…", "matching the CLI's behaviour", "across common install shapes". If the detail doesn't change whether a user cares, cut it.
-- **Lead with what the user can do, not the mechanism.** The reader cares about the capability, not how it works under the hood. Do not explain LAN vs WAN, TLS handshakes, IPC, the daemon-relay topology, or any internal concept the user has not asked about. "Self-hosted relays can use a different TLS setting for the public endpoint" — not "Self-hosted relays support a separate TLS setting for the public endpoint, so the daemon can reach the relay over the LAN while the phone reaches it over the public secure address." If a feature genuinely needs background to be understood, it belongs in product docs, with a one-line teaser in the changelog.
-- **Lead with the outcome.** "Windows: agents launch reliably from npm `.cmd` shims…" is better than "Windows: agents launch reliably across common install shapes. Claude, Codex, and OpenCode now start correctly…".
+- **Stop after identifying the change.** Do not explain LAN/WAN topology, TLS handshakes, IPC, or other architecture in a changelog bullet. Put necessary background in product docs.
 - **Attribution follows the split.** When you split a dense bullet, move each PR/author to the bullet it belongs to. Never duplicate the same PR across multiple bullets.
 
 ## Changelog attribution
@@ -567,6 +577,7 @@ Each beta entry records what its testers receive. Promotion produces the single
 ### Beta release
 
 - [ ] The resolved release source is the intended commit (default `origin/main`) and its existing CI is green
+- [ ] Every PR in the release range has been opened, and its full description and every linked issue have been read before drafting the changelog
 - [ ] Add a new `CHANGELOG.md` entry for this beta (heading `## X.Y.Z-beta.N - YYYY-MM-DD`), review it against the changelog policy, get approval, and commit it before cutting the release
 - [ ] The diff from the previous stable to the resolved release source is classified as patch or minor, with the target version and rationale approved
 - [ ] Release preparation stayed local until the approved release command pushed the complete branch and tag
@@ -587,6 +598,7 @@ Each beta entry records what its testers receive. Promotion produces the single
 - [ ] Run the pre-release sanity check (see above) and address any findings
 - [ ] The diff from the previous stable to the resolved release source is classified as patch or minor, with the target version and rationale approved
 - [ ] The resolved release source is the intended commit (default `origin/main`) and its existing CI is green
+- [ ] Every PR in the release range has been opened, and its full description and every linked issue have been read before drafting the changelog
 - [ ] Ensure the approved release inputs are committed locally and the git worktree is clean before running any release command
 - [ ] Ensure local `npm run typecheck` passes on that exact commit before running any release command
 - [ ] Update `CHANGELOG.md` with user-facing release notes (features, fixes — not refactors). Promotion replaces every `## X.Y.Z-beta.N` entry in the series with one `## X.Y.Z - YYYY-MM-DD` entry covering the full release

docs/terminal-performance.md

@@ -24,7 +24,9 @@ Terminal frames share the daemon main event loop with all agent traffic. The `ev
 - **Coalesced output carries the LAST chunk's revision.** Snapshot replay dedup (`replayTerminalOutputAfterSnapshot`) skips buffered output with `revision <= replayRevision`; a merged batch with a lower revision would be wrongly skipped (lost output).
 - **The input-mode tracker runs once per process boundary, not per hop.** The worker owns the authoritative tracker; the daemon caches the replay preamble from `getTerminalState` responses and `snapshotReady` messages. Do not reintroduce a per-chunk `feed()` on the daemon main loop.
 - **Snapshot catch-up is backpressure-gated.** A stream falls back to a full snapshot only when `outputBytesSinceSnapshot > MAX_TERMINAL_OUTPUT_FRAME_BYTES` (256KB) **and** the client transport reports `bufferedAmount > MAX_CLIENT_BUFFERED_BYTES` (4MB). A client that keeps draining streams continuously, no matter how much output is produced. Before this gate existed, every 256KB of build output dropped a frame and forced a full JSON cell-grid snapshot (~200k objects across IPC) — the historical source of spiky lag and GC hitches.
+- **Plugin daemon sessions report IPC queue bytes.** Their virtual socket increments `bufferedAmount` before `process.send` and decrements it only in the send callback. Text and binary frames share that ordered queue, so the normal snapshot catch-up and physical high-water gates remain valid for server-side plugin SDK traffic.
 - **Client output writes are not serialized per frame.** The emulator runtime drains contiguous plain writes straight into xterm (which buffers internally). Only barrier ops (`clear`, `snapshot`, `suppressInput` writes) wait — behind a zero-length sentinel write — so resets can't interleave with in-flight output.
+- **Retained terminal tabs in the focused workspace keep their streams.** Hidden mounted terminals continue applying output, so switching tabs does not resubscribe or request a fresh snapshot. The retained-panel LRU bounds the number of live streams; terminals in an unfocused workspace detach.
 - **Terminal size has one daemon-owned claimant.** Focus and direct interaction send a `claim`; later geometry changes from that connection send `update`. A claim transfers ownership even when the dimensions are unchanged, while an update from any other connection is ignored. This lets an owning pane follow splits and keyboard insets without allowing an idle phone or browser to steal the PTY size.
 
 ## Measuring

docs/testing.md

@@ -124,6 +124,16 @@ PASEO_DESKTOP_SMOKE_ARTIFACT_DIR=/tmp/paseo-desktop-smoke \
 npm run build:desktop -- --publish never --linux --x64 --dir
 ```
 
+### Undeclared peer dependencies break app.asar
+
+electron-builder packs `node_modules` by walking declared production `dependencies`. A package that imports something it only lists as a `peerDependency` resolves fine in this hoisted workspace, passes every test, and then throws `ERR_MODULE_NOT_FOUND` inside `app.asar` — killing the desktop daemon at startup. That shipped twice from `@replit/codemirror-lang-*` grammars, which are interactive editor extensions published as if they were bare parsers.
+
+The packaged smoke catches it but only runs when a PR touches the `desktop` filter in `.github/ci-paths.yml`. Both offenders landed under `packages/highlight/**`, which maps to `sdk`.
+
+`packages/highlight/src/__tests__/dependency-closure.test.ts` replicates the packer's traversal statically and runs with the normal unit tests. It is scoped to `@getpaseo/highlight` on purpose: that tree is small and pure, so the check is exact. Running the same walk over `@getpaseo/server` produces dozens of false positives from optional dependencies loaded behind `try`/`catch`.
+
+Prefer a `@lezer/*` grammar. When a language only ships inside an editor extension, vendor the grammar into `packages/highlight/src/<lang>/` — see `svelte/`, `nix/`, and `csharp/`.
+
 ### Desktop browser regression
 
 The desktop browser E2E launches an isolated real daemon, Metro, and Electron app. It forces workspace LRU eviction to reparent the original tab and replace its guest `WebContents`, then makes one MCP call each for tab listing, snapshot, and click against that original browser id. A final MCP wait proves the real target page received the click.
@@ -176,9 +186,13 @@ Test suites in this repo are heavy. Running them in bulk freezes the machine, es
 - For full-suite confidence, push to CI and check GitHub Actions.
 - Never run the full Playwright E2E suite locally — defer whole-suite verification to CI. Targeted Playwright specs are allowed when you changed or need to prove that specific flow.
 - App Playwright shares one warmed Metro server per run and gives every Playwright worker its own isolated daemon and `PASEO_HOME`. Spec files run concurrently without exposing one file's projects, agents, terminals, history, or provider configuration to another worker; tests within a file remain together so file-level setup is not repeated.
+- Playwright specs that exercise only the daemon import `daemonTest` from the shared fixtures so they do not create a browser context or page.
 - Helpers that create projects or workspaces own those records until cleanup. Their clients remove the daemon project on close, and an automatic fixture fails any test that still leaks a project record. Deleting only the temporary directory is not cleanup. Agent helpers pass the intended `workspaceId` through to agent creation; they never infer ownership from `cwd`.
 - Tests whose subject is daemon-global state, such as an empty history or daemon restart, start a dedicated host explicitly. Filenames and directories describe product behavior, never execution order or isolation mechanics.
 - Global setup accepts Metro as ready only when `/status` returns `packager-status:running`, then fetches the document's scripts so the cold bundle compilation finishes before Playwright's per-test timeout starts. A generic TCP listener is not sufficient readiness evidence. The browser suite uses direct local daemon connections and does not start a relay.
+- The app Playwright harness boots on Windows as well as POSIX. Spawn Node entrypoints through `process.execPath`, not `npx` or `node_modules/.bin` shims: Node refuses to spawn `.cmd` or `.bat` without `shell: true`, and shell mode concatenates argv without escaping and sends kill signals to `cmd.exe` instead of the real child.
+- Teardown kills the process tree, because a Windows signal reaches only the direct child and leaves forked workers holding the listening port.
+- The `asdf`-backed local Elixir relay stays POSIX-only, so the `relay-deployment` Playwright project is unavailable on Windows.
 
 ## Pull-request test routing
 

docs/timeline-sync.md

@@ -5,6 +5,9 @@ Agent chat delivery has two paths:
 1. **Live stream** — `agent_stream` WebSocket messages for immediacy. These may be delta-shaped lifecycle updates.
 2. **Authoritative history** — `fetch_agent_timeline_request` for correctness. This always returns full projected timeline items, never lifecycle deltas.
 
+The daemon keeps canonical rows only for its runtime. Provider history is the durable transcript
+authority and repopulates those rows when an agent resumes.
+
 The invariants are:
 
 > A continuously subscribed client applies every committed row in order. Opening or resuming an
@@ -12,7 +15,7 @@ The invariants are:
 > through backward pagination.
 
 Tool output is bounded before it enters either delivery path. Canonical shell tool output is sliced
-to 64 KiB, and the same bounded item is used for durable timeline rows and live stream events.
+to 64 KiB, and the same bounded item is used for runtime timeline rows and live stream events.
 Provider history hydration applies the same rule so reopening an agent cannot restore an oversized
 tool payload.
 
@@ -42,6 +45,16 @@ Initialization timeouts guard lack of catch-up progress, not the full multi-page
 Opening or resuming an agent fetches one bounded latest tail page. Older history remains
 user-driven by scrolling upward.
 
+A failed catch-up or subscription reconcile retries on its own, doubling from 1s to a 30s ceiling.
+A fixed 1s retry turned a persistent daemon-side refusal — a Codex thread that already has an active
+writer, say — into a request and a log line every second on an idle app. The delay resets on success,
+reconnect, delivery-mode change, and visibility change, so recovery is still immediate once the
+condition clears. Republishing the same visibility set is a no-op and never bypasses backoff.
+
+Background retries are silent. The retry the user presses in the sync-error callout is a fallible
+user action and owns its pending state: `retrying` is a status the sync model publishes, not a React
+boolean, so the button reports in-flight and the callout returns to `error` when the attempt fails.
+
 Reaching the history-start threshold loads one older page and preserves the visible content anchor.
 Cursor progress does not trigger another page. The user must leave and return to the threshold unless
 the anchored page still leaves the viewport at history start, as with short or compacted content; in
@@ -85,16 +98,22 @@ recomposition while the runtime still owns the same directory snapshot and timel
 Removing the host from the registry is the destructive boundary: it stops the runtime and clears the
 session and host-scoped setup state together.
 
-The durable replica cache is a display cache, not a synchronization checkpoint. Its timeline record
-contains only the focused `agentId` and a truncated item tail. It never persists a cursor, epoch,
-older-history availability, authority status, or sync generation because those facts would describe
-the complete source dataset rather than the truncated display dataset.
+The durable replica cache persists synchronization authority only when it can store the complete
+current canonical window losslessly. The stored range describes those exact items: `startSeq` drives
+older pagination and `endSeq` drives forward catch-up. Restore paints the items immediately, requests
+`after endSeq`, and requests `before startSeq` when the user loads older history.
 
-Restoring that cache produces a painted timeline: the items may render immediately, but the first
-daemon timeline request is still `tail`. A successful tail response atomically establishes canonical
-items, range, and older-history availability. Live rows received between cache paint and that tail
-response stay in the separate live head, do not advance a cursor or trigger gap recovery, and are
-reconciled with the authoritative tail and subsequent catch-up.
+The first resume request is bounded. If it reports more newer history, fetch one latest bounded tail
+instead of replaying every missed page. Live gap recovery still pages forward until current.
+
+If the canonical window exceeds the cache item limit, contains a discontiguous retained range, has a
+live head, or includes presentation data the cache cannot encode losslessly, persistence drops the
+range and keeps a display-only tail. Restore then uses the ordinary bounded `tail` bootstrap. Never
+slice items while retaining the pre-slice range; that falsely certifies discarded source rows.
+
+Live rows received between cache paint and catch-up stay in the separate live head and reconcile with
+the authoritative range through the existing forward-page path. The cache does not persist sync
+generation or unreconciled local submissions.
 
 Every daemon-derived live item carries its timeline epoch and sequence position. Bootstrap
 replacement keeps only positioned rows newer than the page it installs, while unresolved local
@@ -167,6 +186,11 @@ foreground control ownership remains a separate daemon concern. Cancellation req
 with that record rather than in a React component, so an old request cannot clear a newer one. Submissions
 remain a separate pre-turn registry and retire on canonical acknowledgement.
 
+Canonical turns and visible responses are different boundaries. System-injected prompts are absent from
+the Paseo timeline, so one visible response can span several canonical turns without a user message
+between them. Layout and copy group that response together; lifecycle, timing, tool sequences, and exact
+fork positions retain the canonical `turnId` boundaries.
+
 The compatibility boundary for older daemons is snapshot normalization: running/idle status becomes an
 anonymous active turn or idle state once, and downstream code consumes the same activity shape. The app
 does not combine anonymous lifecycle events, timestamps, timeline rows, and resume coverage to infer a

docs/unistyles.md

@@ -335,15 +335,16 @@ If we ever need to avoid the transition entirely, store at least the theme prefe
 
 ## Runtime Theme Patching For User Preferences
 
-Appearance settings (UI/mono font family, font sizes, syntax-highlight theme) are applied by patching every registered theme at runtime with `UnistylesRuntime.updateTheme(name, updater)` — not by threading preference reads through components. `applyAppearance` in `packages/app/src/screens/settings/appearance/apply-appearance.ts` runs from a `ProvidersWrapper` effect on settings load/change and loops all six theme keys, returning `{ ...theme, fontFamily, fontSize, lineHeight, colors.syntax }`.
+Appearance settings (theme, UI/mono font family, font sizes, syntax-highlight theme) are owned by `packages/app/src/appearance`. Its provider subscribes once and synchronizes Unistyles when settings or plugin contributions change. `applyAppearance` patches every key in `REGISTERED_THEMES`, returning `{ ...theme, fontFamily, fontSize, lineHeight, colors.syntax }`.
 
 This works without `useUnistyles()` because every consumer already reads these tokens through `StyleSheet.create((theme) => …)` (or the `withUnistyles`/`uniProps` path for the markdown renderer), so patching the theme repaints tracked views through the native ShadowRegistry with no React re-render.
 
 Gotchas:
 
 - **Patch all themes, not just the active one.** The active theme can change and adaptive mode can flip light/dark; patching every key keeps the active key current and makes ordering vs `setTheme`/`setAdaptiveThemes` irrelevant. The effect depends on the settings values (not on `theme`), so it cannot loop.
+- **The reserved plugin keys are the exception to that ordering.** A plugin-contributed theme is rebuilt from its palette, which discards the appearance patch, so the appearance provider writes the matching light or dark slot before applying font and syntax preferences. See [plugins.md](plugins.md).
 - **Narrow the discriminated union before spreading.** `updateTheme`'s updater returns the theme union; spreading the union widens `colorScheme` to `"light" | "dark"`, which is assignable to neither concrete member. Branch on `t.colorScheme` so each branch spreads a single narrowed theme type (no `as`).
-- **`lineHeight.diff` is the code/diff line-height axis** — it is coupled to the code-font-size control (≈ `codeFontSize * 1.5`). Do NOT use it for prose. Markdown body line-height scales with the UI ramp (`Math.round(theme.fontSize.base * 1.4)`); routing prose through `lineHeight.diff` clips text at small code sizes.
+- **`lineHeight.diff` is the code/diff line-height axis** — it is coupled to the code-font-size control (≈ `codeFontSize * 1.5`). Do NOT use it for prose. Markdown body line-height scales with content size (`Math.round(theme.fontSize.content * 1.4)`); routing prose through `lineHeight.diff` clips text at small code sizes.
 - **High-churn draft values** (live-while-typing in the appearance preview) bypass the theme: apply them as inline styles marked with `inlineUnistylesStyle` so per-keystroke values don't grow the `#unistyles-web` CSS registry.
 - **The app shell uses one `AppearanceStyleBoundary`.** Runtime-patched numeric theme values are baked into Unistyles web classes rather than CSS variables, while parsed/memoized content also does not naturally re-run when appearance tokens change. The boundary sits below stable runtime providers in `app/_layout.tsx` and remounts the visual shell once. `applyAppearance` patches the active theme before inactive registry entries so its subscribers receive the committed values in the same update. Do not add local appearance keys or nested boundaries.
 - **Dynamic font tokens stay widened.** `fontFamily`, `fontSize`, and `lineHeight` on `commonTheme` are annotated `string`/`number` (not narrowed by `as const`) so the updater's return assigns; the platform default stacks live in `DEFAULT_UI_FONT_STACK` / `DEFAULT_MONO_FONT_STACK`.

extensions/linear/README.md

@@ -1,32 +0,0 @@
-# Linear
-
-The official Linear extension adds `Attach Linear issue` to the composer. Search by issue identifier
-or title, then send a stable issue snapshot to the agent.
-
-Set a personal API key in the daemon environment:
-
-```bash
-export LINEAR_API_KEY="lin_api_..."
-```
-
-Register the extension in `$PASEO_HOME/config.json`:
-
-```json
-{
-  "plugins": {
-    "linear": {
-      "source": "directory",
-      "path": "/absolute/path/to/paseo/extensions/linear"
-    }
-  }
-}
-```
-
-Restart the development daemon after changing its environment or plugin configuration.
-
-The entry point demonstrates the complete attachment-source pattern:
-
-- define one validated search RPC;
-- handle it in the daemon subprocess;
-- register a declarative attachment source for the client;
-- keep credentials and vendor API calls out of the client bundle.

extensions/linear/index.tsx

@@ -1,45 +0,0 @@
-import { z } from "zod";
-import {
-  defineAttachmentSource,
-  defineRpc,
-  type PluginAttachmentSearchPayload,
-  type PluginContext,
-} from "@paseo/plugin";
-import { createLinearIssueSearch } from "./linear";
-
-const SearchPayloadSchema = z.object({
-  items: z.array(
-    z.object({
-      id: z.string(),
-      identifier: z.string(),
-      title: z.string(),
-      subtitle: z.string().optional(),
-      url: z.string().url(),
-      text: z.string(),
-      resourceType: z.string(),
-    }),
-  ),
-});
-
-const searchIssues = defineRpc({
-  name: "issues.search",
-  input: z.object({ query: z.string() }),
-  output: SearchPayloadSchema,
-});
-
-const issueAttachments = defineAttachmentSource({
-  id: "issues",
-  title: "Linear issue",
-  icon: "CircleDot",
-  pickerTitle: "Attach Linear issue",
-  searchPlaceholder: "Search by identifier or title",
-  search: searchIssues,
-});
-
-export default function contribute(plugin: PluginContext) {
-  plugin.handle(searchIssues, async ({ query }): Promise<PluginAttachmentSearchPayload> => {
-    const linear = createLinearIssueSearch({ apiKey: process.env.LINEAR_API_KEY ?? "" });
-    return linear.search(query);
-  });
-  plugin.addAttachmentSource(issueAttachments);
-}

nix/desktop-package.nix

@@ -178,11 +178,6 @@ buildNpmPackage {
         exit 1
       fi
 
-      # Skills directory referenced at runtime by some agents
-      if [ -d skills ]; then
-        cp -a skills $out/share/paseo-desktop/
-      fi
-
       # Hicolor icon for desktop environments
       install -Dm644 packages/desktop/assets/icon.png \
         $out/share/icons/hicolor/512x512/apps/paseo-desktop.png

nix/npm-deps.hash

@@ -1 +1 @@
-sha256-BfRbfVUFaHQEMtDOsWAo49txz+cueRpe/7Qxltg+HSg=
+sha256-gZ6egDKIDiDi9jmDlPamM+or2Tthwr17DbJ+P66t4RE=

package.json

@@ -1,6 +1,6 @@
 {
   "name": "paseo",
-  "version": "0.3.0",
+  "version": "0.6.1",
   "private": true,
   "description": "Paseo: voice-controlled development environment for local AI coding agents",
   "keywords": [
@@ -23,6 +23,7 @@
   "workspaces": [
     "packages/expo-two-way-audio",
     "packages/highlight",
+    "packages/plugin",
     "packages/protocol",
     "packages/client",
     "packages/server",
@@ -46,19 +47,21 @@
     "build": "npm run build --workspaces --if-present",
     "build:highlight": "npm run build --workspace=@getpaseo/highlight",
     "build:highlight:clean": "npm run build:clean --workspace=@getpaseo/highlight",
+    "build:plugin": "npm run build --workspace=@getpaseo/plugin",
+    "build:plugin:clean": "npm run build:clean --workspace=@getpaseo/plugin",
     "build:relay": "npm run build --workspace=@getpaseo/relay",
     "build:relay:clean": "npm run build:clean --workspace=@getpaseo/relay",
     "build:protocol": "npm run build --workspace=@getpaseo/protocol",
     "build:protocol:clean": "npm run build:clean --workspace=@getpaseo/protocol",
     "build:client": "npm run build --workspace=@getpaseo/client",
     "build:client:clean": "npm run build:protocol:clean && npm run build:clean --workspace=@getpaseo/client",
-    "build:server-deps": "concurrently --kill-others-on-fail --names highlight,relay,client --prefix-colors yellow,blue,cyan \"npm run build:highlight\" \"npm run build:relay\" \"npm run build:client\"",
-    "build:server-deps:clean": "npm run build:highlight:clean && npm run build:relay:clean && npm run build:client:clean",
+    "build:server-deps": "npm run build:client && concurrently --kill-others-on-fail --names highlight,plugin,relay --prefix-colors yellow,green,blue \"npm run build:highlight\" \"npm run build:plugin\" \"npm run build:relay\"",
+    "build:server-deps:clean": "npm run build:client:clean && npm run build:highlight:clean && npm run build:plugin:clean && npm run build:relay:clean",
     "build:server": "npm run build:server-deps && npm run build --workspace=@getpaseo/server && npm run build --workspace=@getpaseo/cli",
     "build:server:clean": "npm run build:server-deps:clean && npm run build:clean --workspace=@getpaseo/server && npm run build:clean --workspace=@getpaseo/cli",
     "build:daemon-web-ui": "node scripts/build-daemon-web-ui.mjs",
-    "build:app-deps": "npm run build:highlight && npm run build:client && npm run build --workspace=@getpaseo/expo-two-way-audio",
-    "build:app-deps:clean": "npm run build:highlight:clean && npm run build:client:clean && npm run build --workspace=@getpaseo/expo-two-way-audio",
+    "build:app-deps": "npm run build:highlight && npm run build:client && npm run build:plugin && npm run build --workspace=@getpaseo/expo-two-way-audio",
+    "build:app-deps:clean": "npm run build:highlight:clean && npm run build:client:clean && npm run build:plugin:clean && npm run build --workspace=@getpaseo/expo-two-way-audio",
     "watch:protocol": "npm run watch --workspace=@getpaseo/protocol",
     "watch:client": "tsc -p packages/client/tsconfig.json --watch --preserveWatchOutput",
     "typecheck": "npm run typecheck --workspaces --if-present",
@@ -99,11 +102,11 @@
     "version:all:beta:major": "node scripts/set-release-version.mjs --mode beta-major",
     "version:all:beta:next": "node scripts/set-release-version.mjs --mode beta-next",
     "version:all:promote": "node scripts/set-release-version.mjs --mode promote",
-    "release:check": "npm run release:prepare && npm run typecheck --workspace=@getpaseo/highlight && npm run typecheck --workspace=@getpaseo/relay && npm run typecheck --workspace=@getpaseo/protocol && npm run build:client:clean && npm run typecheck --workspace=@getpaseo/client && npm run build:server:clean && npm run typecheck --workspace=@getpaseo/server && npm run typecheck --workspace=@getpaseo/cli && npm pack --dry-run --workspace=@getpaseo/highlight && npm pack --dry-run --workspace=@getpaseo/relay && npm pack --dry-run --workspace=@getpaseo/protocol && npm pack --dry-run --workspace=@getpaseo/client && npm pack --dry-run --workspace=@getpaseo/server && npm pack --dry-run --workspace=@getpaseo/cli",
-    "release:publish:dry-run": "npm publish --dry-run --workspace=@getpaseo/highlight --access public && npm publish --dry-run --workspace=@getpaseo/relay --access public && npm publish --dry-run --workspace=@getpaseo/protocol --access public && npm publish --dry-run --workspace=@getpaseo/client --access public && npm publish --dry-run --workspace=@getpaseo/server --access public && npm publish --dry-run --workspace=@getpaseo/cli --access public",
-    "release:publish:beta:dry-run": "npm publish --dry-run --workspace=@getpaseo/highlight --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/relay --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/protocol --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/client --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/server --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/cli --access public --tag beta",
-    "release:publish": "npm publish --workspace=@getpaseo/highlight --access public && npm publish --workspace=@getpaseo/relay --access public && npm publish --workspace=@getpaseo/protocol --access public && npm publish --workspace=@getpaseo/client --access public && npm publish --workspace=@getpaseo/server --access public && npm publish --workspace=@getpaseo/cli --access public",
-    "release:publish:beta": "npm publish --workspace=@getpaseo/highlight --access public --tag beta && npm publish --workspace=@getpaseo/relay --access public --tag beta && npm publish --workspace=@getpaseo/protocol --access public --tag beta && npm publish --workspace=@getpaseo/client --access public --tag beta && npm publish --workspace=@getpaseo/server --access public --tag beta && npm publish --workspace=@getpaseo/cli --access public --tag beta",
+    "release:check": "npm run release:prepare && npm run typecheck --workspace=@getpaseo/highlight && npm run typecheck --workspace=@getpaseo/relay && npm run typecheck --workspace=@getpaseo/protocol && npm run build:client:clean && npm run typecheck --workspace=@getpaseo/client && npm run typecheck --workspace=@getpaseo/plugin && npm run build:server:clean && npm run typecheck --workspace=@getpaseo/server && npm run typecheck --workspace=@getpaseo/cli && npm pack --dry-run --workspace=@getpaseo/highlight && npm pack --dry-run --workspace=@getpaseo/relay && npm pack --dry-run --workspace=@getpaseo/protocol && npm pack --dry-run --workspace=@getpaseo/client && npm pack --dry-run --workspace=@getpaseo/plugin && npm pack --dry-run --workspace=@getpaseo/server && npm pack --dry-run --workspace=@getpaseo/cli",
+    "release:publish:dry-run": "npm publish --dry-run --workspace=@getpaseo/highlight --access public && npm publish --dry-run --workspace=@getpaseo/relay --access public && npm publish --dry-run --workspace=@getpaseo/protocol --access public && npm publish --dry-run --workspace=@getpaseo/client --access public && npm publish --dry-run --workspace=@getpaseo/plugin --access public && npm publish --dry-run --workspace=@getpaseo/server --access public && npm publish --dry-run --workspace=@getpaseo/cli --access public",
+    "release:publish:beta:dry-run": "npm publish --dry-run --workspace=@getpaseo/highlight --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/relay --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/protocol --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/client --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/plugin --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/server --access public --tag beta && npm publish --dry-run --workspace=@getpaseo/cli --access public --tag beta",
+    "release:publish": "npm publish --workspace=@getpaseo/highlight --access public && npm publish --workspace=@getpaseo/relay --access public && npm publish --workspace=@getpaseo/protocol --access public && npm publish --workspace=@getpaseo/client --access public && npm publish --workspace=@getpaseo/plugin --access public && npm publish --workspace=@getpaseo/server --access public && npm publish --workspace=@getpaseo/cli --access public",
+    "release:publish:beta": "npm publish --workspace=@getpaseo/highlight --access public --tag beta && npm publish --workspace=@getpaseo/relay --access public --tag beta && npm publish --workspace=@getpaseo/protocol --access public --tag beta && npm publish --workspace=@getpaseo/client --access public --tag beta && npm publish --workspace=@getpaseo/plugin --access public --tag beta && npm publish --workspace=@getpaseo/server --access public --tag beta && npm publish --workspace=@getpaseo/cli --access public --tag beta",
     "release:push": "node scripts/push-current-release-tag.mjs",
     "release:beta:patch": "npm run release:check && npm run version:all:beta:patch && npm run release:publish:beta && npm run release:push",
     "release:beta:minor": "npm run release:check && npm run version:all:beta:minor && npm run release:publish:beta && npm run release:push",

packages/app/app.config.js

@@ -1,8 +1,10 @@
 const fs = require("node:fs");
 const path = require("node:path");
 const pkg = require("./package.json");
+const withAndroidAsyncStorageSize = require("./plugins/with-android-async-storage-size");
 const withAndroidProfileable = require("./plugins/with-android-profileable");
 const withFdroidAutolinking = require("./plugins/with-fdroid-autolinking");
+const withPasteInput = require("./plugins/with-paste-input");
 const { getNativeReleaseVersion } = require("./native-release-version");
 const appVariant = process.env.APP_VARIANT ?? "production";
 const isFdroidBuild = process.env.PASEO_FDROID_BUILD === "1";
@@ -138,6 +140,8 @@ export default {
     },
     plugins: [
       "expo-router",
+      withPasteInput,
+      [withAndroidAsyncStorageSize, 64],
       ...buildProfile.cameraPlugins,
       [
         "expo-splash-screen",

packages/app/e2e/browser/acp-provider-catalog.spec.ts

@@ -11,8 +11,8 @@ import {
 } from "../support/helpers/settings";
 
 const ACP_PROVIDER = {
-  id: "hermes",
-  name: "Hermes",
+  id: "minimax-code",
+  name: "MiniMax Code",
 };
 
 interface ProviderCatalogDaemonClient {
@@ -22,7 +22,7 @@ interface ProviderCatalogDaemonClient {
 }
 
 test.describe("ACP provider catalog", () => {
-  test("adds a catalog provider from settings", async ({ page }) => {
+  test("adds MiniMax Code from settings", async ({ page }) => {
     const client = await connectDaemonClient<ProviderCatalogDaemonClient>({
       clientIdPrefix: "provider-catalog-e2e",
     });

packages/app/e2e/browser/add-changed-file-to-chat.spec.ts

@@ -2,6 +2,7 @@ import { mkdir, writeFile } from "node:fs/promises";
 import path from "node:path";
 import { test, expect, type Page } from "../support/fixtures";
 import { seedMockAgentWorkspace, openAgentRoute } from "../support/helpers/mock-agent";
+import { openChangesPanel } from "../support/helpers/workspace-tabs";
 
 function visibleComposer(page: Page) {
   return page.locator("textarea[data-composer-input]").filter({ visible: true }).first();
@@ -31,8 +32,7 @@ test("adds a changed file to the focused chat without replacing its composer dra
     await expect(agentComposer).toBeEditable({ timeout: 30_000 });
     await agentComposer.fill("Preserve this thought");
 
-    await page.getByRole("button", { name: "Open explorer" }).click();
-    await page.getByTestId("explorer-tab-changes").click();
+    await openChangesPanel(page);
     const changedFile = page.getByText("changed file.ts", { exact: true }).first();
     await expect(changedFile).toBeVisible({ timeout: 30_000 });
     await page.getByTestId("diff-file-0-toggle").click({ button: "right" });

packages/app/e2e/browser/add-project-flow.spec.ts

@@ -13,6 +13,7 @@ import {
   expectAddProjectPage,
   expectNewWorkspaceForAddedProject,
   openAddProjectFlow,
+  openAddProjectHostSelection,
 } from "../support/helpers/add-project-flow";
 import { gotoAppShell } from "../support/helpers/app";
 import {
@@ -65,52 +66,21 @@ async function expectProjectHasNoWorkspaces(projectId: string): Promise<void> {
 test.describe("Add Project command-center flow", () => {
   test.describe.configure({ timeout: 180_000 });
 
-  test("a single connected host opens directly on method selection", async ({ page }) => {
+  test("method selection shows the daemon's available project sources without search", async ({
+    page,
+  }) => {
     await gotoAppShell(page);
 
     await openAddProjectFlow(page);
 
     await expect(addProjectFlowMethod(page, "directory-search")).toBeVisible();
+    await expect(addProjectFlowMethod(page, "github")).toContainText("Clone from GitHub");
+    await expect(addProjectFlowMethod(page, "new-directory")).toContainText("New directory");
+    await expect(addProjectFlowInput(page)).toHaveCount(0);
+    await expect(addProjectFlow(page).getByRole("textbox")).toHaveCount(0);
     await expect(page.getByTestId("add-project-flow-page-host")).toHaveCount(0);
   });
 
-  test("the back arrow, search input, and result glyph share one left edge", async ({ page }) => {
-    await gotoAppShell(page);
-    await openAddProjectFlow(page);
-
-    await page.keyboard.press("Enter");
-    await expectAddProjectPage(page, "directory-search");
-    await addProjectFlowInput(page).fill("/tmp");
-
-    const backGlyph = addProjectFlowBack(page).locator("svg");
-    const resultGlyph = addProjectFlow(page)
-      .locator('[data-testid^="add-project-flow-path-"]')
-      .first()
-      .locator("svg");
-    await expect(resultGlyph).toBeVisible();
-
-    const [backBox, inputBox, resultBox, titleBox, resultsBox, footerBox] = await Promise.all([
-      backGlyph.boundingBox(),
-      addProjectFlowInput(page).boundingBox(),
-      resultGlyph.boundingBox(),
-      page.getByTestId("add-project-flow-title").boundingBox(),
-      page.getByTestId("add-project-flow-results").boundingBox(),
-      page.getByTestId("add-project-flow-footer").boundingBox(),
-    ]);
-    expect(backBox).not.toBeNull();
-    expect(inputBox).not.toBeNull();
-    expect(resultBox).not.toBeNull();
-    expect(titleBox).not.toBeNull();
-    expect(resultsBox).not.toBeNull();
-    expect(footerBox).not.toBeNull();
-    if (!backBox || !inputBox || !resultBox || !titleBox || !resultsBox || !footerBox) return;
-
-    expect(Math.abs(backBox.x - inputBox.x)).toBeLessThanOrEqual(2);
-    expect(Math.abs(resultBox.x - inputBox.x)).toBeLessThanOrEqual(2);
-    expect(titleBox.height).toBeLessThanOrEqual(24);
-    expect(resultsBox.y + resultsBox.height).toBeLessThanOrEqual(footerBox.y + 1);
-  });
-
   test("an offline extra host neither appears nor forces host selection", async ({ page }) => {
     await gotoAppShell(page);
     await addOfflineHostAndReload(page, {
@@ -146,7 +116,7 @@ test.describe("Add Project command-center flow", () => {
         serverId: SECONDARY_HOST_ID,
         endpoint: `localhost:${secondaryHost.port}`,
       });
-      await openAddProjectFlow(page, "host");
+      await openAddProjectHostSelection(page);
 
       await page.keyboard.press("ArrowDown");
       await page.keyboard.press("Enter");
@@ -155,7 +125,7 @@ test.describe("Add Project command-center flow", () => {
       await expect(addProjectFlow(page)).toContainText(SECONDARY_HOST_LABEL);
     });
 
-    test("Escape and Back restore page input and active selection before closing at the root", async ({
+    test("Escape and Back restore searchable page input before closing at the root", async ({
       page,
     }) => {
       await gotoAppShell(page);
@@ -168,21 +138,20 @@ test.describe("Add Project command-center flow", () => {
         serverId: SECONDARY_HOST_ID,
         endpoint: `localhost:${secondaryHost.port}`,
       });
-      await openAddProjectFlow(page, "host");
+      await openAddProjectHostSelection(page);
 
       await addProjectFlowInput(page).fill("o");
       await page.keyboard.press("ArrowDown");
       await page.keyboard.press("Enter");
       await expectAddProjectPage(page, "method");
 
-      await addProjectFlowInput(page).fill("new");
-      await page.keyboard.press("Enter");
+      await chooseAddProjectMethod(page, "new-directory");
       await expectAddProjectPage(page, "new-directory-parent");
       await page.keyboard.press("Escape");
 
       await expectAddProjectPage(page, "method");
-      await expect(addProjectFlowInput(page)).toHaveValue("new");
-      await page.keyboard.press("Enter");
+      await expect(addProjectFlowInput(page)).toHaveCount(0);
+      await chooseAddProjectMethod(page, "new-directory");
       await expectAddProjectPage(page, "new-directory-parent");
       await addProjectFlowBack(page).click();
 
@@ -216,7 +185,7 @@ test.describe("Add Project command-center flow", () => {
           serverId: SECONDARY_HOST_ID,
           endpoint: `localhost:${secondaryHost.port}`,
         });
-        await openAddProjectFlow(page, "host");
+        await openAddProjectHostSelection(page);
         await addProjectFlowHost(page, SECONDARY_HOST_ID).click();
         await expectAddProjectPage(page, "method");
 
@@ -271,14 +240,6 @@ test.describe("Add Project command-center flow", () => {
     await expectProjectHasNoWorkspaces(projectId);
   });
 
-  test("the current daemon advertises Clone from GitHub and New directory", async ({ page }) => {
-    await gotoAppShell(page);
-    await openAddProjectFlow(page);
-
-    await expect(addProjectFlowMethod(page, "github")).toContainText("Clone from GitHub");
-    await expect(addProjectFlowMethod(page, "new-directory")).toContainText("New directory");
-  });
-
   test("a complete repository URL remains selectable without a GitHub search result", async ({
     page,
   }) => {

packages/app/e2e/browser/agent-consecutive-turns.spec.ts

@@ -442,21 +442,40 @@ function formatFrames(frames: TurnFrame[], centers: number | number[]): string {
     .join("\n");
 }
 
+function submittedRowContentTop(frame: TurnFrame | undefined): number | undefined {
+  const viewportTop = frame?.userRow.rect?.top;
+  const scrollTop = frame?.scroll.scrollTop;
+  return viewportTop == null || scrollTop == null ? undefined : viewportTop + scrollTop;
+}
+
 function expectRowContinuity(frames: TurnFrame[]): void {
   const first = frames.findIndex((frame) => hasPaintedLayout(frame.userRow));
-  const baseline = frames[first]?.userRow.rect?.top;
+  const baselineFrame = frames[first];
+  const baselineContentTop = submittedRowContentTop(baselineFrame);
   const disappeared = frames.findIndex(
     (frame, index) => index > first && !hasPaintedLayout(frame.userRow),
   );
   const reappeared = frames.findIndex(
     (frame, index) => index > disappeared && disappeared >= 0 && hasPaintedLayout(frame.userRow),
   );
-  const shifted = frames.findIndex(
-    (frame, index) =>
+  const contentShifted = frames.findIndex((frame, index) => {
+    const contentTop = submittedRowContentTop(frame);
+    return (
       index > first &&
       hasPaintedLayout(frame.userRow) &&
-      baseline !== undefined &&
-      Math.abs((frame.userRow.rect?.top ?? baseline) - baseline) > 1,
+      baselineContentTop !== undefined &&
+      contentTop !== undefined &&
+      Math.abs(contentTop - baselineContentTop) > 1
+    );
+  });
+  const viewportChanged = frames.findIndex(
+    (frame, index) =>
+      index > first &&
+      [
+        geometryChanged(frame.scroll.rect?.top, baselineFrame?.scroll.rect?.top),
+        geometryChanged(frame.scroll.rect?.height, baselineFrame?.scroll.rect?.height),
+        geometryChanged(frame.scroll.clientHeight, baselineFrame?.scroll.clientHeight),
+      ].some(Boolean),
   );
   const attachmentMissing = frames.findIndex(
     (frame, index) => index >= first && !hasPaintedLayout(frame.attachment),
@@ -468,16 +487,18 @@ function expectRowContinuity(frames: TurnFrame[]): void {
           `submitted row disappeared for ${Math.max(0, reappeared - disappeared)} recorded frames (${Math.max(0, (frames[reappeared]?.at ?? frames.at(-1)?.at ?? 0) - frames[disappeared].at).toFixed(1)}ms) before authoritative recovery`,
         ]
       : []),
-    ...(shifted >= 0
+    ...(contentShifted >= 0
       ? [
-          `submitted row moved from ${baseline?.toFixed(1)} to ${frames[shifted].userRow.rect?.top.toFixed(1)}`,
+          `submitted row moved in scroll content from ${baselineContentTop?.toFixed(1)} to ${submittedRowContentTop(frames[contentShifted])?.toFixed(1)}`,
         ]
       : []),
+    ...(viewportChanged >= 0 ? ["scroll viewport geometry changed"] : []),
     ...(attachmentMissing >= 0 ? ["submitted image attachment left the painted layout"] : []),
   ];
   let failure = Math.max(0, first);
   if (attachmentMissing >= 0) failure = attachmentMissing;
-  if (shifted >= 0) failure = shifted;
+  if (viewportChanged >= 0) failure = viewportChanged;
+  if (contentShifted >= 0) failure = contentShifted;
   if (disappeared >= 0) failure = disappeared;
   expect(
     violations,
@@ -726,7 +747,7 @@ async function recordDelayedRunningTransition(
   gate.holdNextAgentStreamEvent("turn_started");
   gate.holdNextAgentUpdate(agent.agentId, "running");
   gate.holdNextServerMessage("send_agent_message_response");
-  gate.setAgentStreamItemSuppressed("assistant_message", true);
+  gate.setAgentStreamSuppressed(true);
   await recordTurnFrames(page, prompt);
   await installActivityContinuityOracle(page, prompt);
   await submitMessage(page, prompt);
@@ -779,13 +800,13 @@ async function recordDelayedRunningTransition(
     const checkpoints = await readActivityContinuityOracle(page);
     expectActivityContinuity(checkpoints);
     expect(elapsedBeforeAuthoritative).toBe(0);
-    gate.setAgentStreamItemSuppressed("assistant_message", false);
+    gate.setAgentStreamSuppressed(false);
     return { frames, checkpoints };
   } finally {
     if (!turnReleased) gate.releaseHeldAgentStreamEvent("turn_started");
     if (!snapshotReleased) gate.releaseHeldAgentUpdate(agent.agentId, "running");
     if (!responseReleased) gate.releaseHeldServerMessage("send_agent_message_response");
-    gate.setAgentStreamItemSuppressed("assistant_message", false);
+    gate.setAgentStreamSuppressed(false);
   }
 }
 
@@ -814,6 +835,8 @@ test("keeps the first prompt of a new agent in place through authoritative hydra
     await expectComposerVisible(page);
 
     const prompt = "Delay synthetic user message by 300ms.";
+    gate.holdNextServerMessage("fetch_agent_timeline_response");
+    gate.setAgentStreamEventSuppressed("timeline", true);
     await attachImageFromMenu(page, FIRST_PROMPT_IMAGE);
     await expectAttachmentPill(page, "composer-image-attachment-pill");
     await recordTurnFrames(page, prompt);
@@ -821,7 +844,10 @@ test("keeps the first prompt of a new agent in place through authoritative hydra
 
     const submittedRow = page.getByTestId("user-message").filter({ hasText: prompt }).first();
     await expect(submittedRow).toBeVisible();
-    await gate.waitForServerMessage("fetch_agent_timeline_response");
+    await gate.waitForHeldServerMessage("fetch_agent_timeline_response");
+    gate.truncateHeldTimelineAfterLast("user_message");
+    gate.releaseHeldServerMessage("fetch_agent_timeline_response");
+    await expect(submittedRow.getByTestId("rewind-menu-trigger")).toBeVisible();
     await recordPaintsFor(page, 80);
     expectAtomicFirstPromptTransition(await stopTurnFrameRecording(page));
   } finally {

packages/app/e2e/browser/agent-message-rewind.spec.ts

@@ -1,34 +1,13 @@
 import type { Locator } from "@playwright/test";
 import { expect, test, type Page } from "../support/fixtures";
 import { expectAgentIdle, expectAgentReadyToInterrupt } from "../support/helpers/agent-stream";
-import {
-  cancelAgent,
-  expectComposerDraft,
-  expectComposerVisible,
-  submitMessage,
-} from "../support/helpers/composer";
+import { cancelAgent, expectComposerVisible, submitMessage } from "../support/helpers/composer";
 import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
 
-async function completeSubmittedTurn(
-  page: Page,
-  agent: Awaited<ReturnType<typeof seedMockAgentWorkspace>>,
-  prompt: string,
-): Promise<Locator> {
-  await submitMessage(page, prompt);
-  const userMessage = page.getByTestId("user-message").filter({ hasText: prompt });
-  await expect(userMessage).toBeVisible();
-
-  const finish = await agent.client.waitForFinish(agent.agentId, 30_000);
-  expect(finish.status).toBe("idle");
-  await expect(page.getByText("(end of synthetic stream)", { exact: true })).toBeVisible();
-  await expect(userMessage).toHaveAttribute("aria-busy", "false");
-  return userMessage;
-}
-
 async function rewindConversation(page: Page, userMessage: Locator, prompt: string): Promise<void> {
   await userMessage.getByText(prompt, { exact: true }).hover();
   await userMessage.getByRole("button", { name: "Rewind to this message", exact: true }).click();
-  await page.getByRole("button", { name: "Rewind conversation", exact: true }).click();
+  await page.getByRole("menuitem", { name: "Rewind conversation", exact: true }).click();
 }
 
 async function expectTurnCompletesNormally(
@@ -42,29 +21,6 @@ async function expectTurnCompletesNormally(
 }
 
 test.describe("Agent message rewind", () => {
-  test("rewinds a submitted prompt and restores it to the composer", async ({ page }) => {
-    test.setTimeout(90_000);
-    const prompt = "Restore this submitted prompt after rewind.";
-    const agent = await seedMockAgentWorkspace({
-      repoPrefix: "message-rewind-roundtrip-",
-      title: "Message rewind roundtrip",
-      model: "ten-second-stream",
-    });
-
-    try {
-      await openAgentRoute(page, agent);
-      await expectComposerVisible(page);
-      const userMessage = await completeSubmittedTurn(page, agent, prompt);
-
-      await rewindConversation(page, userMessage, prompt);
-
-      await expect(page.getByText("(end of synthetic stream)", { exact: true })).toHaveCount(0);
-      await expectComposerDraft(page, prompt);
-    } finally {
-      await agent.cleanup();
-    }
-  });
-
   test("keeps a pre-acknowledgement turn running after rewind is rejected", async ({ page }) => {
     test.setTimeout(90_000);
     const prompt = "Delay synthetic user message by 2000ms.";

packages/app/e2e/browser/agent-message-submission.spec.ts

@@ -24,12 +24,23 @@ import {
 } from "../support/helpers/composer";
 import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
 import { seedWorkspace } from "../support/helpers/seed-client";
-import { waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+import {
+  createAgentTabFromMenu,
+  waitForWorkspaceTabsVisible,
+} from "../support/helpers/workspace-tabs";
 import { getServerId } from "../support/helpers/server-id";
 import { buildHostWorkspaceRoute } from "@/utils/host-routes";
+import { WORKSPACE_DECK_MAX_MOUNTED_WORKSPACES } from "@/screens/workspace/workspace-deck-retention";
 import { delayBrowserAgentCreatedStatus } from "../support/helpers/new-workspace";
 import { installDaemonWebSocketGate } from "../support/helpers/daemon-websocket-gate";
-import { selectModel } from "../support/helpers/app";
+import { gotoAppShell, openSettings, selectModel } from "../support/helpers/app";
+import { observeTimelineSubscriptions } from "../support/helpers/timeline-delivery";
+import {
+  expectResumeOverflowFallsBackToOneTail,
+  rememberTimelineRequestCounts,
+} from "../support/helpers/timeline-resume";
+import { workspaceDeckEntryLocator } from "../support/helpers/workspace-ui";
+import { expectInFlightForkAvailable } from "../support/helpers/assistant-fork";
 import {
   scrollTimelineToNewestLoadedEdge,
   scrollTimelineUntilOlderHistoryIsReachable,
@@ -257,14 +268,20 @@ async function submitMessageThatWillBeRejected(page: Page, prompt: string): Prom
 
 async function expectRejectedSubmissionRestored(
   page: Page,
-  input: { prompt: string; errorMessage: string },
+  input: { prompt: string; errorMessage: string; preservesActiveTurn?: boolean },
 ): Promise<void> {
-  await expect(page.getByText(input.errorMessage)).toBeVisible({ timeout: 30_000 });
+  await expect(page.getByRole("alert").filter({ hasText: input.errorMessage })).toBeVisible({
+    timeout: 30_000,
+  });
   await expectComposerDraft(page, input.prompt);
   await expectComposerEditable(page);
   await expectAttachmentPill(page, "composer-image-attachment-pill");
-  await expect(page.getByRole("button", { name: "Send message" })).toBeEnabled();
   await expect(page.getByTestId("user-message").filter({ hasText: input.prompt })).toHaveCount(0);
+  if (input.preservesActiveTurn) {
+    await expect(page.getByTestId("turn-working-indicator")).toBeVisible();
+    return;
+  }
+  await expect(page.getByRole("button", { name: "Send message" })).toBeEnabled();
   await expect(page.getByTestId("turn-working-indicator")).toHaveCount(0);
 }
 
@@ -277,6 +294,77 @@ async function retryRestoredSubmission(page: Page, prompt: string): Promise<void
   await expect(page.getByTestId("composer-image-attachment-pill")).toHaveCount(0);
 }
 
+async function configureSteerInSettings(page: Page): Promise<void> {
+  const modifier = process.platform === "darwin" ? "Meta" : "Control";
+  await page.keyboard.press(`${modifier}+Comma`);
+  await expect(page).toHaveURL(/\/settings\/general$/);
+  await selectSteerInSettings(page);
+}
+
+async function selectSteerInSettings(page: Page): Promise<void> {
+  await selectSendBehaviorInSettings(page, "Steer", "steer");
+}
+
+/** Steer is the default, so the interrupt path only gets exercised by opting back into it. */
+async function configureInterruptInSettings(page: Page): Promise<void> {
+  const modifier = process.platform === "darwin" ? "Meta" : "Control";
+  await page.keyboard.press(`${modifier}+Comma`);
+  await expect(page).toHaveURL(/\/settings\/general$/);
+  await selectSendBehaviorInSettings(page, "Interrupt", "interrupt");
+}
+
+async function selectSendBehaviorInSettings(
+  page: Page,
+  behaviorLabel: string,
+  stored: string,
+): Promise<void> {
+  await page.getByRole("button", { name: /^Default send: / }).click();
+  await page.getByRole("menuitem", { name: behaviorLabel, exact: true }).click();
+  await expect
+    .poll(async () => {
+      const raw = await page.evaluate(() => localStorage.getItem("@paseo:app-settings"));
+      return raw ? (JSON.parse(raw) as { sendBehavior?: unknown }).sendBehavior : null;
+    })
+    .toBe(stored);
+}
+
+async function replaySteeredSleepTurnInBrowser(
+  page: Page,
+  testInfo: { workerIndex: number },
+  shape: "claude" | "codex",
+): Promise<void> {
+  const gate = await installDaemonWebSocketGate(page);
+  gate.holdNextShellToolCall("completed");
+  await gotoAppShell(page);
+  await openSettings(page);
+  await selectSteerInSettings(page);
+  const agent = await startRunningMockAgent(page, {
+    prefix: `steer-replay-${shape}-${testInfo.workerIndex}-`,
+    model: "ten-second-stream",
+    prompt: `Replay a ${shape}-shaped foreground shell tool call while the user steers this turn.`,
+  });
+  try {
+    await expect(page.getByTestId("tool-call-badge").last()).toBeVisible({ timeout: 30_000 });
+    await expectComposerVisible(page);
+    await submitMessage(page, "hello");
+
+    await expect(page.getByText("hello", { exact: true })).toHaveCount(1);
+    await expect(page.getByRole("button", { name: /^Worked for/ })).toHaveCount(0);
+    await expectInFlightForkAvailable(page);
+
+    await gate.waitForHeldServerMessage();
+    gate.releaseHeldServerMessage();
+    await agent.client.waitForFinish(agent.agentId, 30_000);
+
+    await expect(page.getByText("hello", { exact: true })).toHaveCount(1);
+    await expect(page.getByRole("button", { name: /^Worked for/ })).toHaveCount(1);
+    await expect(page.getByRole("button", { name: "Fork chat" }).last()).toBeVisible();
+  } finally {
+    gate.restore();
+    await agent.cleanup();
+  }
+}
+
 async function queueMessage(page: Page, prompt: string): Promise<void> {
   await fillComposerDraft(page, prompt);
   await sendDraftToQueue(page);
@@ -336,6 +424,66 @@ async function expectInterruptedTurnOrderAfterReconnect(
   }
 }
 
+async function expectHiddenStreamingSubmissionOrderAfterWorkspaceEviction(
+  page: Page,
+  testInfo: { workerIndex: number },
+): Promise<void> {
+  const subscriptions = observeTimelineSubscriptions(page);
+  const gate = await installDaemonWebSocketGate(page);
+  const target = await seedMockAgentWorkspace({
+    repoPrefix: `submission-hidden-stream-${testInfo.workerIndex}-`,
+    title: "Hidden streaming submission",
+    model: "ten-second-stream",
+  });
+  const evictionAgents = await Promise.all(
+    Array.from({ length: WORKSPACE_DECK_MAX_MOUNTED_WORKSPACES }, (_unused, index) =>
+      seedMockAgentWorkspace({
+        repoPrefix: `submission-workspace-eviction-${testInfo.workerIndex}-${index}-`,
+        title: `Workspace eviction ${index + 1}`,
+      }),
+    ),
+  );
+  const prompt = "Keep this hidden image prompt before its streaming output.";
+  const targetDeckEntry = workspaceDeckEntryLocator(page, getServerId(), target.workspaceId);
+
+  try {
+    await openAgentRoute(page, target);
+    await expectComposerVisible(page);
+    await subscriptions.waitForSubscribedAgents([target.agentId]);
+
+    const userMessageCount = gate.getAgentStreamItemCount("user_message");
+    gate.setAgentStreamSuppressed(true);
+    const promptRow = await submitMessageWithImage(page, prompt);
+    await gate.waitForAgentStreamItem("user_message", userMessageCount + 1);
+
+    for (const evictionAgent of evictionAgents) {
+      await openAgentRoute(page, evictionAgent);
+      await expectComposerVisible(page);
+    }
+    await expect(targetDeckEntry).toHaveCount(0);
+    await subscriptions.waitForSubscribedAgents([
+      evictionAgents[WORKSPACE_DECK_MAX_MOUNTED_WORKSPACES - 1]!.agentId,
+    ]);
+    gate.setAgentStreamSuppressed(false);
+
+    await target.client.waitForFinish(target.agentId, 30_000);
+    const requestsBeforeReturn = rememberTimelineRequestCounts(gate);
+    await openAgentRoute(page, target);
+    await expectComposerVisible(page);
+    await subscriptions.waitForSubscribedAgents([target.agentId]);
+
+    const response = page.getByText("(end of synthetic stream)", { exact: true }).last();
+    await expect(promptRow).toBeVisible();
+    await expect(response).toBeVisible();
+    await expectRenderedBefore(promptRow, response);
+    expectResumeOverflowFallsBackToOneTail(gate, requestsBeforeReturn);
+  } finally {
+    gate.setAgentStreamSuppressed(false);
+    gate.restore();
+    await Promise.all([...evictionAgents.map((agent) => agent.cleanup()), target.cleanup()]);
+  }
+}
+
 async function expectCompletedSubmissionClearsAfterMissedRunningTransition(
   page: Page,
   testInfo: { workerIndex: number },
@@ -627,7 +775,7 @@ async function expectRenderedBefore(first: Locator, second: Locator): Promise<vo
 async function openWorkspaceDraft(page: Page, workspaceId: string): Promise<void> {
   await page.goto(buildHostWorkspaceRoute(getServerId(), workspaceId));
   await waitForWorkspaceTabsVisible(page);
-  await page.getByTestId("workspace-new-agent-tab-inline").click();
+  await createAgentTabFromMenu(page);
   await expectComposerVisible(page);
 }
 
@@ -801,9 +949,9 @@ test.describe("Agent message submission", () => {
 
       gate.holdNextClientRequest("send_agent_message_request");
       await fillComposerDraft(page, "Replace the running turn without duplicating its action.");
-      await expect(
-        page.getByRole("button", { name: "Send and interrupt", exact: true }),
-      ).toHaveCount(1);
+      await expect(page.getByRole("button", { name: "Send and steer", exact: true })).toHaveCount(
+        1,
+      );
       await expect(page.getByRole("button", { name: "Stop agent", exact: true })).toHaveCount(0);
       await expect(page.getByRole("button", { name: "Interrupt agent", exact: true })).toHaveCount(
         0,
@@ -963,6 +1111,115 @@ test.describe("Agent message submission", () => {
     await retryRestoredSubmission(page, prompt);
   });
 
+  test("restores an ambiguous Steer failure without retrying or interrupting", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(120_000);
+    const gate = await installDaemonWebSocketGate(page);
+    const prompt = "Restore this ambiguous Steer submission.";
+    const agent = await startRunningMockAgent(page, {
+      prefix: `steer-ambiguous-${testInfo.workerIndex}-`,
+      model: "one-minute-stream",
+      prompt: "Keep this turn active while a Steer request fails.",
+      featureValues: { mockSteerAmbiguousFailures: 1 },
+    });
+    try {
+      await configureSteerInSettings(page);
+      await page.goBack();
+      await expectComposerVisible(page);
+      await expectAgentReadyToInterrupt(page);
+      const sendsBefore = gate.getClientRequestCount("send_agent_message_request");
+      const cancelsBefore = gate.getClientRequestCount("cancel_agent_request");
+
+      await submitMessageThatWillBeRejected(page, prompt);
+      await expectRejectedSubmissionRestored(page, {
+        prompt,
+        errorMessage: "Requested mock steer transport failure",
+        preservesActiveTurn: true,
+      });
+
+      expect(gate.getClientRequestCount("send_agent_message_request")).toBe(sendsBefore + 1);
+      expect(gate.getClientRequestCount("cancel_agent_request")).toBe(cancelsBefore);
+      expect(gate.getClientRequests("send_agent_message_request").at(-1)).toMatchObject({
+        text: prompt,
+        activeTurnBehavior: "steer",
+      });
+    } finally {
+      gate.restore();
+      await agent.cleanup();
+    }
+  });
+
+  test("keeps an optimistic Steer prompt inside the active turn before acknowledgement", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(120_000);
+    const gate = await installDaemonWebSocketGate(page);
+    const agent = await startRunningMockAgent(page, {
+      prefix: `steer-optimistic-${testInfo.workerIndex}-`,
+      model: "one-minute-stream",
+      prompt: "Keep this turn active while the user steers it.",
+    });
+    try {
+      await configureSteerInSettings(page);
+      await page.goBack();
+      await expectComposerVisible(page);
+      await expectAgentReadyToInterrupt(page);
+      gate.holdNextServerMessage("send_agent_message_response");
+      await submitMessage(page, "hello");
+      await gate.waitForHeldServerMessage("send_agent_message_response");
+      await expect(page.getByText("hello", { exact: true })).toHaveCount(1);
+      await expect(page.getByText(/^Worked for/)).toHaveCount(0);
+      gate.releaseHeldServerMessage("send_agent_message_response");
+      await expect(page.getByText("hello", { exact: true })).toHaveCount(1);
+    } finally {
+      gate.restore();
+      await agent.cleanup();
+    }
+  });
+
+  test("sends interrupt behavior on the wire when the user opts out of steering", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(120_000);
+    const gate = await gateNextAgentMessage(page);
+    const agent = await startRunningMockAgent(page, {
+      prefix: `interrupt-submission-${testInfo.workerIndex}-`,
+      model: "one-minute-stream",
+      prompt: "Keep this turn active until the user interrupts it.",
+    });
+    try {
+      await configureInterruptInSettings(page);
+      await page.goBack();
+      await expectComposerVisible(page);
+      await expectAgentReadyToInterrupt(page);
+
+      const prompt = "Interrupt the running turn.";
+      await fillComposerDraft(page, prompt);
+      await expect(
+        page.getByRole("button", { name: "Send and interrupt", exact: true }),
+      ).toHaveCount(1);
+      await composerLocator(page).press("Enter");
+
+      const request = await gate.waitForRequest();
+      expect(request.activeTurnBehavior).toBe("interrupt");
+      gate.accept();
+      await expect(page.getByTestId("user-message").filter({ hasText: prompt })).toHaveCount(1);
+    } finally {
+      await agent.cleanup();
+    }
+  });
+
+  test("replays Claude-shaped steering inside one active turn", async ({ page }, testInfo) => {
+    test.setTimeout(120_000);
+    await replaySteeredSleepTurnInBrowser(page, testInfo, "claude");
+  });
+
+  test("replays Codex-shaped steering inside one active turn", async ({ page }, testInfo) => {
+    test.setTimeout(120_000);
+    await replaySteeredSleepTurnInBrowser(page, testInfo, "codex");
+  });
+
   test("restores overlapping queued sends when their connection fails", async ({
     page,
   }, testInfo) => {
@@ -1013,6 +1270,13 @@ test.describe("Agent message submission", () => {
     await expectInterruptedTurnOrderAfterReconnect(page, testInfo);
   });
 
+  test("keeps a streaming hidden submission before its output after workspace eviction", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(240_000);
+    await expectHiddenStreamingSubmissionOrderAfterWorkspaceEviction(page, testInfo);
+  });
+
   test("clears an attachment-only submission when canonical history arrives after a missed running transition", async ({
     page,
   }, testInfo) => {

packages/app/e2e/browser/agent-profile-modeless-preferences.spec.ts

@@ -0,0 +1,240 @@
+import { expect, test, type Page } from "../support/fixtures";
+import path from "node:path";
+import type { FormPreferences } from "@/create-agent-preferences/preferences";
+import {
+  applyProfileFromPicker,
+  drillIntoProvider,
+  openModelPicker,
+  seedAgentProfiles,
+  seedModelProvider,
+} from "../support/helpers/agent-profiles";
+import { gotoAppShell } from "../support/helpers/app";
+import { daemonWsRoutePattern } from "../support/helpers/daemon-port";
+import {
+  openGlobalNewWorkspaceComposer,
+  selectNewWorkspaceProject,
+  submitNewWorkspacePrompt,
+} from "../support/helpers/new-workspace";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { waitForSidebarHydration } from "../support/helpers/workspace-ui";
+
+const CREATE_AGENT_PREFERENCES_KEY = "@paseo:create-agent-preferences";
+const MODELESS_PROVIDER = "modeless-profile-e2e";
+const MODELESS_MODEL = "pi-profile-model";
+
+type WebSocketMessage = string | Buffer;
+
+interface CreateAgentRequestMessage {
+  type: "create_agent_request";
+  config?: {
+    provider?: unknown;
+    modeId?: unknown;
+    featureValues?: unknown;
+  };
+}
+
+function getSessionMessage(message: WebSocketMessage): Record<string, unknown> | null {
+  const rawMessage = typeof message === "string" ? message : message.toString("utf8");
+  try {
+    const envelope = JSON.parse(rawMessage) as { type?: unknown; message?: unknown };
+    return envelope.type === "session" && envelope.message && typeof envelope.message === "object"
+      ? (envelope.message as Record<string, unknown>)
+      : null;
+  } catch {
+    return null;
+  }
+}
+
+async function recordAndBlockCreateAgentRequest(page: Page): Promise<{
+  waitForRequest(): Promise<CreateAgentRequestMessage>;
+}> {
+  let resolveRequest: ((message: CreateAgentRequestMessage) => void) | null = null;
+  const request = new Promise<CreateAgentRequestMessage>((resolve) => {
+    resolveRequest = resolve;
+  });
+
+  await page.routeWebSocket(daemonWsRoutePattern(), (ws) => {
+    const server = ws.connectToServer();
+    ws.onMessage((message) => {
+      const sessionMessage = getSessionMessage(message);
+      if (sessionMessage?.type === "create_agent_request") {
+        resolveRequest?.(sessionMessage as unknown as CreateAgentRequestMessage);
+        return;
+      }
+      server.send(message);
+    });
+    server.onMessage((message) => ws.send(message));
+  });
+
+  return { waitForRequest: () => request };
+}
+
+async function seedPoisonedModelessPreference(page: Page): Promise<void> {
+  await page.addInitScript(
+    ({ preferencesKey, provider, model }) => {
+      localStorage.setItem(
+        preferencesKey,
+        JSON.stringify({
+          provider,
+          providerPreferences: {
+            [provider]: { model, mode: "full-access" },
+            mock: { model: "ten-second-stream", mode: "load-test" },
+          },
+        } satisfies FormPreferences),
+      );
+    },
+    {
+      preferencesKey: CREATE_AGENT_PREFERENCES_KEY,
+      provider: MODELESS_PROVIDER,
+      model: MODELESS_MODEL,
+    },
+  );
+}
+
+async function readProviderModePreference(page: Page, provider: string): Promise<unknown> {
+  return page.evaluate(
+    ({ preferencesKey, providerId }) => {
+      const raw = localStorage.getItem(preferencesKey);
+      if (!raw) return null;
+      const preferences = JSON.parse(raw) as FormPreferences;
+      return preferences.providerPreferences?.[providerId]?.mode ?? null;
+    },
+    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY, providerId: provider },
+  );
+}
+
+async function readProviderFeaturePreferences(page: Page, provider: string): Promise<unknown> {
+  return page.evaluate(
+    ({ preferencesKey, providerId }) => {
+      const raw = localStorage.getItem(preferencesKey);
+      if (!raw) return null;
+      const preferences = JSON.parse(raw) as FormPreferences;
+      return preferences.providerPreferences?.[providerId]?.featureValues ?? null;
+    },
+    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY, providerId: provider },
+  );
+}
+
+test.describe("Agent profiles repair modeless provider preferences", () => {
+  test.describe.configure({ timeout: 240_000 });
+
+  test("applying a profile heals a stale mode before returning to Pi", async ({ page }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "profile-modeless-preferences-" });
+    const profiles = await seedAgentProfiles([
+      {
+        id: "agent_profile_modeless_repair",
+        name: "Approval work",
+        provider: "mock",
+        model: "one-minute-stream",
+        modeId: "approval-test",
+      },
+    ]);
+    const provider = await seedModelProvider({
+      id: MODELESS_PROVIDER,
+      label: "Pi profile test",
+      extends: "pi",
+      command: [process.execPath, path.resolve("e2e/fixtures/fake-pi-rpc.mjs")],
+      models: [
+        {
+          id: MODELESS_MODEL,
+          label: "Pi profile model",
+          description: "Modeless profile regression model",
+        },
+      ],
+    });
+    await seedPoisonedModelessPreference(page);
+    const createAgentRecorder = await recordAndBlockCreateAgentRequest(page);
+
+    try {
+      await gotoAppShell(page);
+      await waitForSidebarHydration(page);
+      await openGlobalNewWorkspaceComposer(page);
+      await selectNewWorkspaceProject(page, {
+        projectKey: workspace.projectKey,
+        projectDisplayName: workspace.projectDisplayName,
+      });
+
+      await openModelPicker(page);
+      await applyProfileFromPicker(page, "Approval work");
+
+      await expect
+        .poll(() => readProviderModePreference(page, MODELESS_PROVIDER), { timeout: 10_000 })
+        .toBeNull();
+      await expect
+        .poll(() => readProviderModePreference(page, "mock"), { timeout: 10_000 })
+        .toBe("approval-test");
+
+      await openModelPicker(page);
+      await drillIntoProvider(page, MODELESS_PROVIDER);
+      await page.getByRole("button", { name: /Pi profile model/ }).click();
+
+      await submitNewWorkspacePrompt(page, "Create a Pi agent after repairing preferences.");
+      const createAgentRequest = await createAgentRecorder.waitForRequest();
+      expect(createAgentRequest).toMatchObject({
+        config: { provider: MODELESS_PROVIDER },
+      });
+      expect(createAgentRequest.config).not.toHaveProperty("modeId");
+    } finally {
+      await workspace.cleanup();
+      await provider.restore();
+      await profiles.restore();
+    }
+  });
+
+  test("cross-provider profile features survive immediate workspace creation", async ({ page }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "profile-feature-transition-" });
+    const featureProviderId = "profile-feature-e2e";
+    const featureModelId = "claude-opus-5";
+    const profiles = await seedAgentProfiles([
+      {
+        id: "agent_profile_feature_transition",
+        name: "Fast profile",
+        provider: featureProviderId,
+        model: featureModelId,
+        modeId: "bypassPermissions",
+        featureValues: { fast_mode: true },
+      },
+    ]);
+    const provider = await seedModelProvider({
+      id: featureProviderId,
+      label: "Profile feature test",
+      models: [
+        {
+          id: featureModelId,
+          label: "Profile feature model",
+          description: "Cross-provider feature regression model",
+        },
+      ],
+    });
+    const createAgentRecorder = await recordAndBlockCreateAgentRequest(page);
+
+    try {
+      await gotoAppShell(page);
+      await waitForSidebarHydration(page);
+      await openGlobalNewWorkspaceComposer(page);
+      await selectNewWorkspaceProject(page, {
+        projectKey: workspace.projectKey,
+        projectDisplayName: workspace.projectDisplayName,
+      });
+
+      await openModelPicker(page);
+      await applyProfileFromPicker(page, "Fast profile");
+      await expect
+        .poll(() => readProviderFeaturePreferences(page, featureProviderId), { timeout: 10_000 })
+        .toEqual({ fast_mode: true });
+      await submitNewWorkspacePrompt(page, "Create an agent with profile features.");
+
+      const createAgentRequest = await createAgentRecorder.waitForRequest();
+      expect(createAgentRequest).toMatchObject({
+        config: {
+          provider: featureProviderId,
+          featureValues: { fast_mode: true },
+        },
+      });
+    } finally {
+      await workspace.cleanup();
+      await provider.restore();
+      await profiles.restore();
+    }
+  });
+});

packages/app/e2e/browser/agent-profiles-picker.spec.ts

@@ -0,0 +1,118 @@
+import { expect, test } from "../support/fixtures";
+import {
+  applyProfileFromPicker,
+  closeModelPicker,
+  expectComposerDoesNotName,
+  expectAgentProfilesEmptyPrompt,
+  expectProfileEditTooltip,
+  expectComposerMode,
+  expectComposerModel,
+  expectModelRowSelected,
+  expectProfileEditIsPencilOnly,
+  expectProfileVisibleForProvider,
+  openModelPicker,
+  openAgentProfilesFromEmptyPrompt,
+  seedAgentProfiles,
+} from "../support/helpers/agent-profiles";
+import { expectWorkspaceAgentConfiguration } from "../support/helpers/command-center-agent-controls";
+import { expectComposerVisible } from "../support/helpers/composer";
+import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
+
+const PROFILE = {
+  id: "agent_profile_e2e_ui_work",
+  name: "UI work",
+  icon: "🎨",
+  provider: "mock",
+  model: "one-minute-stream",
+  modeId: "approval-test",
+  notes: "Use for UI work.",
+};
+
+const PROFILE_SUMMARY = "Mock Load Test · One minute stream · Approval test";
+
+test.describe("Agent profiles in the model picker", () => {
+  test("an empty host still exposes agent profile settings from the picker", async ({ page }) => {
+    const seed = await seedAgentProfiles([]);
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "agent-profiles-empty-",
+      title: "Agent profiles empty",
+    });
+
+    try {
+      await openAgentRoute(page, workspace);
+      await expectComposerVisible(page);
+      await openModelPicker(page);
+      await expectAgentProfilesEmptyPrompt(page);
+      await openAgentProfilesFromEmptyPrompt(page);
+    } finally {
+      await workspace.cleanup();
+      await seed.restore();
+    }
+  });
+
+  test("applying a pinned profile materializes it into the composer and is then forgotten", async ({
+    page,
+  }) => {
+    const seed = await seedAgentProfiles([PROFILE]);
+    // A live agent is one provider's process, so the profile has to name that
+    // same provider or the picker will not offer it at all.
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "agent-profiles-picker-",
+      title: "Agent profiles picker",
+      model: "ten-second-stream",
+      modeId: "load-test",
+    });
+
+    try {
+      await test.step("the agent starts on its seeded model and mode", async () => {
+        await openAgentRoute(page, workspace);
+        await expectComposerVisible(page);
+        await expectComposerModel(page, "Ten second stream");
+        await expectComposerMode(page, "Load test");
+      });
+
+      await test.step("the sole provider opens directly", async () => {
+        await openModelPicker(page);
+        await expect(page.getByTestId("model-search-input").first()).toBeVisible();
+        await expect(page.getByTestId("sheet-header-back")).toHaveCount(0);
+        await expect(page.locator('[data-testid^="model-provider-"]')).toHaveCount(0);
+        await expectProfileVisibleForProvider(page, {
+          name: PROFILE.name,
+          summary: PROFILE_SUMMARY,
+        });
+        await expectProfileEditIsPencilOnly(page);
+        await expectProfileEditTooltip(page);
+      });
+
+      await test.step("applying it writes its model and mode into the composer", async () => {
+        await applyProfileFromPicker(page, PROFILE.name);
+        await expectComposerModel(page, "One minute stream");
+        await expectComposerMode(page, "Approval test");
+        await expectWorkspaceAgentConfiguration(workspace, {
+          id: workspace.agentId,
+          provider: "mock",
+          model: "one-minute-stream",
+          modeId: "approval-test",
+        });
+      });
+
+      await test.step("the composer names the model, never the profile", async () => {
+        await expectComposerDoesNotName(page, PROFILE.name);
+      });
+
+      await test.step("reopening returns directly to the provider models", async () => {
+        await openModelPicker(page);
+        await expect(page.getByTestId("model-search-input").first()).toBeVisible();
+        await expectProfileVisibleForProvider(page, {
+          name: PROFILE.name,
+          summary: PROFILE_SUMMARY,
+        });
+        await expectModelRowSelected(page, { provider: "mock", modelId: "one-minute-stream" });
+        await closeModelPicker(page);
+      });
+    } finally {
+      await workspace.cleanup();
+      await seed.restore();
+    }
+  });
+});

packages/app/e2e/browser/agent-profiles-settings.spec.ts

@@ -0,0 +1,128 @@
+import { test } from "../support/fixtures";
+import {
+  createAgentProfile,
+  createAgentProfileFromEmptyState,
+  editAgentProfile,
+  expectAgentProfile,
+  expectAgentProfileForm,
+  expectAgentProfileOrder,
+  expectAgentProfileTagsGone,
+  expectHostAgentProfiles,
+  expectNoAgentProfiles,
+  moveAgentProfileUp,
+  openAgentProfileSettings,
+  removeAgentProfile,
+  seedAgentProfiles,
+  stageLegacyFavoritesForHostMigration,
+} from "../support/helpers/agent-profiles";
+
+const MOCK_PROVIDER_LABEL = "Mock Load Test";
+
+test.describe("Agent profiles settings", () => {
+  test("legacy model favourites migrate into provider-and-model-only host profiles", async ({
+    page,
+  }) => {
+    const seed = await seedAgentProfiles([]);
+    const migratedProfile = {
+      id: "legacy_favorite:mock:one-minute-stream",
+      name: "One minute stream",
+      provider: "mock",
+      model: "one-minute-stream",
+    };
+
+    try {
+      await stageLegacyFavoritesForHostMigration(page, [
+        { provider: "mock", modelId: "one-minute-stream" },
+      ]);
+
+      await expectHostAgentProfiles([migratedProfile]);
+      await openAgentProfileSettings(page);
+      await expectAgentProfile(page, {
+        name: migratedProfile.name,
+        tags: [MOCK_PROVIDER_LABEL, "One minute stream"],
+      });
+    } finally {
+      await seed.restore();
+    }
+  });
+
+  test("host owner creates, edits, reorders and removes agent profiles", async ({ page }) => {
+    // The journey owns the list, so it starts from an empty one whatever the
+    // worker ran before, and hands it back untouched.
+    const seed = await seedAgentProfiles([]);
+
+    try {
+      await test.step("the section starts empty", async () => {
+        await openAgentProfileSettings(page);
+        await expectNoAgentProfiles(page);
+      });
+
+      await test.step("create a profile from the modal", async () => {
+        await createAgentProfileFromEmptyState(page, {
+          name: "UI work",
+          provider: MOCK_PROVIDER_LABEL,
+          model: "Ten second stream",
+          thinking: "High",
+          mode: "Approval test",
+          notes: "Use for UI work — components, layout and design tokens.",
+        });
+        await expectAgentProfile(page, {
+          name: "UI work",
+          tags: [MOCK_PROVIDER_LABEL, "Ten second stream", "Approval test", "High"],
+          notes: "Use for UI work — components, layout and design tokens.",
+        });
+      });
+
+      await test.step("editing to a model without thinking levels drops the level", async () => {
+        await editAgentProfile(page, "UI work", {
+          model: "One minute stream",
+          notes: "Now for quick checks.",
+        });
+        await expectAgentProfile(page, {
+          name: "UI work",
+          tags: [MOCK_PROVIDER_LABEL, "One minute stream", "Approval test"],
+          notes: "Now for quick checks.",
+        });
+        await expectAgentProfileTagsGone(page, {
+          name: "UI work",
+          tags: ["Ten second stream", "High"],
+        });
+      });
+
+      await test.step("the edit form reopens on what was stored", async () => {
+        await expectAgentProfileForm(page, "UI work", {
+          provider: MOCK_PROVIDER_LABEL,
+          model: "One minute stream",
+          mode: "Approval test",
+          notes: "Now for quick checks.",
+        });
+      });
+
+      await test.step("a second profile lands below the first", async () => {
+        await createAgentProfile(page, {
+          name: "Deep review",
+          provider: MOCK_PROVIDER_LABEL,
+          model: "Five minute stream",
+        });
+        await expectAgentProfileOrder(page, ["UI work", "Deep review"]);
+      });
+
+      await test.step("reorder puts the second profile on top", async () => {
+        await moveAgentProfileUp(page, "Deep review");
+        await expectAgentProfileOrder(page, ["Deep review", "UI work"]);
+      });
+
+      await test.step("removing a profile confirms by name", async () => {
+        await removeAgentProfile(page, "UI work");
+        await expectAgentProfileOrder(page, ["Deep review"]);
+      });
+
+      await test.step("removing the last profile returns the empty state", async () => {
+        await removeAgentProfile(page, "Deep review");
+        await expectNoAgentProfiles(page);
+      });
+    } finally {
+      await seed.restore();
+    }
+  });
+});

packages/app/e2e/browser/agent-skills-settings.spec.ts

@@ -0,0 +1,215 @@
+import { access } from "node:fs/promises";
+import { expect, test as base } from "../support/fixtures";
+import {
+  answerNextRemovalWarning,
+  cancelSkillSelection,
+  chooseAllSkills,
+  chooseCustomSkills,
+  expectInstalledSkills,
+  expectSavedSelection,
+  expectSkillFile,
+  openAgentSkillsSettings,
+  openSkillSelection,
+  saveSkillSelection,
+  startAgentSkillsSandbox,
+  toggleSkill,
+  type AgentSkillsSandbox,
+} from "../support/helpers/agent-skills";
+import { readVerticalGap, waitForSettledPosition } from "../support/helpers/sheet-layout";
+
+const test = base.extend<{ skills: AgentSkillsSandbox }>({
+  // oxlint-disable-next-line no-empty-pattern -- Playwright requires destructuring for fixture dependency discovery.
+  skills: async ({}, provide) => {
+    const sandbox = await startAgentSkillsSandbox();
+    try {
+      await provide(sandbox);
+    } finally {
+      await sandbox.close();
+    }
+  },
+});
+
+test.describe("Host agent skills", () => {
+  test.describe.configure({ timeout: 120_000 });
+
+  test("installs a custom selection, cancels a draft, then installs all skills", async ({
+    page,
+    skills,
+  }) => {
+    const custom = skills.available.slice(0, 2);
+    const cancelledAddition = skills.available[2]!;
+    await openAgentSkillsSettings(page, skills);
+
+    await openSkillSelection(page);
+    await chooseCustomSkills(page, custom);
+    await saveSkillSelection(page);
+    await expectInstalledSkills(skills, custom);
+    await expectSavedSelection(skills, { mode: "custom", skills: custom });
+
+    await openSkillSelection(page);
+    await toggleSkill(page, cancelledAddition);
+    await cancelSkillSelection(page);
+    await expectInstalledSkills(skills, custom);
+    await expectSavedSelection(skills, { mode: "custom", skills: custom });
+
+    await openSkillSelection(page);
+    await chooseAllSkills(page);
+    await saveSkillSelection(page);
+    await expectSavedSelection(skills, { mode: "all" });
+    await expectInstalledSkills(skills, skills.available);
+  });
+
+  test("dismisses and then accepts removal confirmation", async ({ page, skills }) => {
+    const retained = skills.available.slice(0, 2);
+    const removed = skills.available.slice(2);
+    await skills.install({ mode: "all" });
+    await openAgentSkillsSettings(page, skills);
+    await openSkillSelection(page);
+    await chooseCustomSkills(page, retained);
+
+    const dismissedWarning = answerNextRemovalWarning(page, "dismiss");
+    await saveSkillSelection(page);
+    const dismissedDialog = await dismissedWarning;
+    expect(dismissedDialog.message()).toContain("Remove deselected skills?");
+    for (const skill of removed) expect(dismissedDialog.message()).toContain(skill);
+    await expectInstalledSkills(skills, skills.available);
+    await expectSavedSelection(skills, { mode: "all" });
+
+    const acceptedWarning = answerNextRemovalWarning(page, "accept");
+    await saveSkillSelection(page);
+    expect((await acceptedWarning).message()).toContain(
+      "Anything you added inside those skill folders",
+    );
+    await expectInstalledSkills(skills, retained);
+    await expectSavedSelection(skills, { mode: "custom", skills: retained });
+  });
+
+  test("warns before removing a partial installation", async ({ page, skills }) => {
+    const partial = skills.available.at(-1)!;
+    const retained = skills.available.filter((name) => name !== partial);
+    await skills.install({ mode: "all" });
+    await skills.keepSkillOnlyIn(partial, "claude");
+    await openAgentSkillsSettings(page, skills);
+    await openSkillSelection(page);
+    await chooseCustomSkills(page, retained);
+
+    const warning = answerNextRemovalWarning(page, "dismiss");
+    await saveSkillSelection(page);
+    expect((await warning).message()).toContain(partial);
+    await expectSkillFile(skills, "claude", partial, "SKILL.md");
+    await expectSavedSelection(skills, { mode: "all" });
+  });
+
+  test("warns without deleting user files from an externally restored skill", async ({
+    page,
+    skills,
+  }) => {
+    const retained = skills.available.slice(0, 2);
+    const drift = skills.available[2]!;
+    await skills.install({ mode: "custom", skills: retained });
+    await skills.placeSkillOnDisk(drift, "claude", {
+      "SKILL.md": `# ${drift}\n`,
+      "notes/mine.md": "user notes",
+    });
+    await openAgentSkillsSettings(page, skills);
+    await openSkillSelection(page);
+
+    const warning = answerNextRemovalWarning(page, "dismiss");
+    await saveSkillSelection(page);
+    expect((await warning).message()).toContain(drift);
+    await expectSkillFile(skills, "claude", drift, "notes/mine.md");
+    await expectSavedSelection(skills, { mode: "custom", skills: retained });
+  });
+
+  test("keeps the sheet open with an error when convergence fails", async ({ page, skills }) => {
+    await openAgentSkillsSettings(page, skills);
+    await openSkillSelection(page);
+    await skills.blockAgentsDirectory();
+    await chooseCustomSkills(page, [skills.available[0]!]);
+    await saveSkillSelection(page);
+
+    await expect(
+      page.getByText("Could not save your skill selection.", { exact: true }),
+    ).toBeVisible();
+    await expect(page.getByRole("switch", { name: "All skills", exact: true })).toBeVisible();
+    await skills.unblockAgentsDirectory();
+    await expectSavedSelection(skills, { mode: "all" });
+  });
+
+  test("serializes concurrent saves through the daemon command surface", async ({ skills }) => {
+    const firstSelection = { mode: "custom" as const, skills: [skills.available[0]!] };
+    const secondSelection = { mode: "custom" as const, skills: [skills.available[1]!] };
+
+    const first = skills.client.saveAgentSkillsSelection(firstSelection, skills.available);
+    const second = skills.client.saveAgentSkillsSelection(secondSelection, skills.available);
+    await expect(Promise.all([first, second])).resolves.toMatchObject([
+      { confirmationRequired: null, selection: firstSelection },
+      { confirmationRequired: null, selection: secondSelection },
+    ]);
+
+    await expectInstalledSkills(skills, secondSelection.skills);
+    await expectSavedSelection(skills, secondSelection);
+  });
+
+  test("ignores agent config environment overrides", async ({ skills }) => {
+    await skills.install({ mode: "custom", skills: [skills.available[0]!] });
+
+    await expect(
+      access(`${skills.home}/../ignored-claude-config/skills`).then(
+        () => true,
+        () => false,
+      ),
+    ).resolves.toBe(false);
+    await expect(
+      access(`${skills.home}/../ignored-codex-home/skills`).then(
+        () => true,
+        () => false,
+      ),
+    ).resolves.toBe(false);
+  });
+});
+
+function allSkillsCard(page: Parameters<typeof openSkillSelection>[0]) {
+  return page.getByTestId("skill-selection-all-card");
+}
+
+function skillListCard(page: Parameters<typeof openSkillSelection>[0]) {
+  return page.getByTestId("skill-selection-list-card");
+}
+
+function bundledSkillsLabel(page: Parameters<typeof openSkillSelection>[0]) {
+  return page.getByText("Bundled skills", { exact: true });
+}
+
+test.describe("Agent skills sheet inset", () => {
+  test.describe.configure({ timeout: 120_000 });
+
+  test.describe("compact", () => {
+    test.use({ viewport: { width: 390, height: 844 } });
+
+    test("keeps more space above the bundled skills label than below", async ({ page, skills }) => {
+      await openAgentSkillsSettings(page, skills, { compact: true });
+      await openSkillSelection(page);
+      await waitForSettledPosition(page.getByTestId("skill-selection-sheet"));
+
+      expect({
+        aboveLabel: await readVerticalGap(allSkillsCard(page), bundledSkillsLabel(page)),
+        belowLabel: await readVerticalGap(bundledSkillsLabel(page), skillListCard(page)),
+      }).toEqual({ aboveLabel: 16, belowLabel: 12 });
+    });
+  });
+
+  test.describe("desktop", () => {
+    test.use({ viewport: { width: 1280, height: 900 } });
+
+    test("keeps more space above the bundled skills label than below", async ({ page, skills }) => {
+      await openAgentSkillsSettings(page, skills);
+      await openSkillSelection(page);
+
+      expect({
+        aboveLabel: await readVerticalGap(allSkillsCard(page), bundledSkillsLabel(page)),
+        belowLabel: await readVerticalGap(bundledSkillsLabel(page), skillListCard(page)),
+      }).toEqual({ aboveLabel: 16, belowLabel: 12 });
+    });
+  });
+});

packages/app/e2e/browser/agent-timeline-pagination-old-daemon.spec.ts

@@ -66,7 +66,7 @@ test("does not repeat an assistant block when the current app paginates a publis
         localStorage.setItem("@paseo:daemon-registry", JSON.stringify([seededHost]));
         localStorage.setItem("@paseo:create-agent-preferences", JSON.stringify(preferences));
       },
-      { seededHost: host, preferences: buildCreateAgentPreferences(serverId) },
+      { seededHost: host, preferences: buildCreateAgentPreferences() },
     );
 
     const history = await holdOlderHistoryPages(page, agent, daemon.port);

packages/app/e2e/browser/agent-timeline-pagination.spec.ts

@@ -26,9 +26,41 @@ import {
   expectTimelineViewportAnchoredAfterPrepend,
   userNavigatesTimelineToHistoryStartWithKeyboard,
   userScrollsTimelineToHistoryStart,
+  waitForPersistedCanonicalTimelineRange,
 } from "../support/helpers/timeline-pagination";
+import { trackAgentTimelineRequests } from "../support/helpers/agent-timeline-gate";
+import { seedMockAgentWorkspace } from "../support/helpers/mock-agent";
 
 test.describe("Agent timeline pagination", () => {
+  test("resumes a persisted canonical timeline after its exact end", async ({ page }) => {
+    const prompt = "timeline persisted range resumes without tail replay";
+    const agent = await seedMockAgentWorkspace({
+      repoPrefix: "timeline-persisted-range-",
+      title: "Persisted timeline range",
+      initialPrompt: prompt,
+    });
+    try {
+      await agent.client.waitForFinish(agent.agentId, 15_000);
+      await openAgentTimeline(page, agent);
+      await expectTimelinePromptVisible(page, prompt);
+      const range = await waitForPersistedCanonicalTimelineRange(page, agent.agentId);
+
+      const tracker = await trackAgentTimelineRequests(page, agent.agentId);
+      await page.reload();
+      await expectTimelinePromptVisible(page, prompt);
+
+      const request = await tracker.nextRequest();
+      expect(request).toEqual({
+        direction: "after",
+        cursor: { epoch: range.epoch, seq: range.endSeq },
+      });
+      await tracker.waitForResponse();
+      expect(tracker.requests().some((entry) => entry.direction === "tail")).toBe(false);
+    } finally {
+      await agent.cleanup();
+    }
+  });
+
   test("keeps a fixed history-start gutter before, during, and after pagination", async ({
     page,
   }) => {
@@ -59,14 +91,16 @@ test.describe("Agent timeline pagination", () => {
       await expectTimelinePromptVisible(page, agent.newestPrompt);
       await expectTimelinePromptCentered(page, agent.newestPrompt);
       await expectTimelinePromptNotMounted(page, agent.oldestPrompt);
+      await expectTimelinePromptNotMounted(page, agent.initialTailOldestPrompt);
 
       await userScrollsTimelineToHistoryStart(page);
+      await expectTimelinePromptVisible(page, agent.initialTailOldestPrompt);
       await history.expectRequestedPages(1);
       history.releasePage(1);
       await history.expectSettledWithRequestedPages(1);
-      await expectTimelinePromptCentered(page, agent.firstOlderPagePrompt);
 
       await userScrollsTimelineToHistoryStart(page);
+      await expectTimelinePromptCentered(page, agent.firstOlderPagePrompt);
       await history.expectRequestedPages(2);
     } finally {
       await agent.cleanup();

packages/app/e2e/browser/agent-timeline-resume.spec.ts

@@ -13,9 +13,10 @@ import {
 import {
   commitTimelineTurnsWhileDisconnected,
   disconnectViewedTimeline,
-  expectOneTailCheckWithoutCatchUpPages,
+  expectOneResumeCheckWithoutTail,
+  expectResumeOverflowFallsBackToOneTail,
   rememberTimelineRequestCounts,
-  restoreViewedTimelineWithHeldTail,
+  restoreViewedTimelineWithHeldResponse,
 } from "../support/helpers/timeline-resume";
 
 test.describe("Agent timeline resume", () => {
@@ -29,11 +30,11 @@ test.describe("Agent timeline resume", () => {
       const requests = rememberTimelineRequestCounts(gate);
       await disconnectViewedTimeline(page, gate);
       const background = await commitTimelineTurnsWhileDisconnected(agent, 24);
-      await restoreViewedTimelineWithHeldTail(page, gate);
+      await restoreViewedTimelineWithHeldResponse(page, gate);
 
       await expectTimelinePromptVisible(page, background.newestPrompt);
       await expectTimelinePromptNotMounted(page, background.firstPrompt);
-      expectOneTailCheckWithoutCatchUpPages(gate, requests);
+      expectResumeOverflowFallsBackToOneTail(gate, requests);
       await scrollTimelineUntilOlderHistoryIsReachable(page, background.firstPrompt);
     } finally {
       gate.restore();
@@ -53,10 +54,10 @@ test.describe("Agent timeline resume", () => {
       const requests = rememberTimelineRequestCounts(gate);
 
       await disconnectViewedTimeline(page, gate);
-      await restoreViewedTimelineWithHeldTail(page, gate);
+      await restoreViewedTimelineWithHeldResponse(page, gate);
 
       await expectTimelinePresentationUnchanged(page, presentation);
-      expectOneTailCheckWithoutCatchUpPages(gate, requests);
+      expectOneResumeCheckWithoutTail(gate, requests);
     } finally {
       gate.restore();
       await agent.cleanup();

packages/app/e2e/browser/agent-timeline-sync-retry.spec.ts

@@ -0,0 +1,87 @@
+import { expect, test } from "../support/fixtures";
+import { installDaemonWebSocketGate } from "../support/helpers/daemon-websocket-gate";
+import {
+  expectTimelinePromptVisible,
+  openAgentTimeline,
+  seedLongMockAgentTimeline,
+} from "../support/helpers/timeline-pagination";
+import {
+  disconnectViewedTimeline,
+  restoreViewedTimelineWithHeldResponse,
+} from "../support/helpers/timeline-resume";
+
+test.describe("Agent timeline sync retry", () => {
+  test("a refused catch-up offers a retry that reports pending, failure, and recovery", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(120_000);
+    const gate = await installDaemonWebSocketGate(page);
+    const agent = await seedLongMockAgentTimeline({ turns: 2 });
+    try {
+      await openAgentTimeline(page, agent);
+      await expectTimelinePromptVisible(page, agent.newestPrompt);
+
+      gate.failTimelineResponses(agent.agentId);
+      await disconnectViewedTimeline(page, gate);
+      await restoreViewedTimelineWithHeldResponse(page, gate);
+
+      const callout = page.getByTestId("agent-timeline-sync-error");
+      const retry = page.getByTestId("agent-timeline-sync-retry");
+      await expect(callout).toBeVisible({ timeout: 30_000 });
+      await expect(retry).toBeEnabled();
+      await page.screenshot({ path: testInfo.outputPath("1-sync-error.png"), fullPage: true });
+
+      gate.holdTimelineResponses(agent.agentId);
+      await retry.click();
+      await gate.waitForHeldTimelineResponse();
+      await expect(retry).toBeDisabled();
+      await expect(retry).toHaveText("Retrying…");
+      await page.screenshot({ path: testInfo.outputPath("2-retry-pending.png"), fullPage: true });
+
+      gate.releaseHeldTimelineResponses();
+      await expect(retry).toBeEnabled({ timeout: 30_000 });
+      await expect(callout).toBeVisible();
+      await page.screenshot({ path: testInfo.outputPath("3-retry-failed.png"), fullPage: true });
+
+      gate.allowTimelineResponses();
+      await retry.click();
+      await expect(callout).toBeHidden({ timeout: 30_000 });
+      await expectTimelinePromptVisible(page, agent.newestPrompt);
+      await page.screenshot({ path: testInfo.outputPath("4-recovered.png"), fullPage: true });
+    } finally {
+      gate.restore();
+      await agent.cleanup();
+    }
+  });
+
+  test("a refused first load offers a retry instead of a dead panel", async ({
+    page,
+  }, testInfo) => {
+    test.setTimeout(120_000);
+    const gate = await installDaemonWebSocketGate(page);
+    const agent = await seedLongMockAgentTimeline({ turns: 2 });
+    try {
+      gate.failTimelineResponses(agent.agentId);
+      await openAgentTimeline(page, agent);
+
+      const loadError = page.getByTestId("agent-load-error");
+      const retry = page.getByTestId("agent-load-error-retry");
+      await expect(loadError).toBeVisible({ timeout: 60_000 });
+      await expect(loadError).toContainText("already has an active writer");
+      await page.screenshot({ path: testInfo.outputPath("5-load-error.png"), fullPage: true });
+
+      await retry.click();
+      await expect(loadError).toBeVisible({ timeout: 30_000 });
+      await page.screenshot({ path: testInfo.outputPath("6-load-retry-failed.png") });
+
+      gate.allowTimelineResponses();
+      await retry.click();
+      await expect(loadError).toBeHidden({ timeout: 30_000 });
+      await expectTimelinePromptVisible(page, agent.newestPrompt);
+      await page.screenshot({ path: testInfo.outputPath("7-load-recovered.png"), fullPage: true });
+    } finally {
+      gate.restore();
+      await agent.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/appearance-theme-picker.spec.ts

@@ -1,4 +1,7 @@
 import { expect, test } from "../support/fixtures";
+import { gotoAppShell } from "../support/helpers/app";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { getServerId } from "../support/helpers/server-id";
 import { openSettingsSection } from "../support/helpers/settings";
 
 test("shows Pure black in the appearance picker", async ({ page }, testInfo) => {
@@ -15,9 +18,36 @@ test("shows Pure black in the appearance picker", async ({ page }, testInfo) =>
   });
 });
 
+test("keeps the selected workspace visible in Pure black", async ({ page }, testInfo) => {
+  const workspace = await seedWorkspace({
+    repoPrefix: "pure-black-selected-workspace-",
+    title: "Selected workspace",
+  });
+
+  try {
+    await page.addInitScript(() => {
+      localStorage.setItem("@paseo:app-settings", JSON.stringify({ theme: "pureBlack" }));
+    });
+    await gotoAppShell(page);
+
+    const row = page.getByTestId(`sidebar-workspace-row-${getServerId()}:${workspace.workspaceId}`);
+    await expect(row).toBeVisible({ timeout: 30_000 });
+    await row.click();
+
+    await expect(row).toHaveAttribute("aria-selected", "true");
+    await expect(row).toHaveCSS("background-color", "rgb(17, 17, 17)");
+    await page.screenshot({
+      path: testInfo.outputPath("pure-black-selected-workspace.png"),
+      fullPage: true,
+    });
+  } finally {
+    await workspace.cleanup();
+  }
+});
+
 test("applies the interface font size to settings text", async ({ page }) => {
   await page.addInitScript(() => {
-    localStorage.setItem("@paseo:app-settings", JSON.stringify({ uiFontSize: 24 }));
+    localStorage.setItem("@paseo:app-settings", JSON.stringify({ uiBaseFontSize: 21 }));
   });
   await page.goto("/settings");
   await expect(page.getByTestId("settings-sidebar")).toBeVisible();
@@ -26,11 +56,14 @@ test("applies the interface font size to settings text", async ({ page }) => {
   const sectionTitle = page.getByText("Theme", { exact: true }).first();
   await expect(sectionTitle).toHaveCSS("font-size", "18px");
 
-  const fontSizeInput = page.getByLabel("Interface font size");
-  await expect(fontSizeInput).toHaveValue("24");
-  await fontSizeInput.fill("12");
-  await fontSizeInput.press("Tab");
+  const interfaceSizeInput = page.getByLabel("Interface font size");
+  const contentSizeInput = page.getByLabel("Content font size");
+  await expect(interfaceSizeInput).toHaveValue("21");
+  await expect(contentSizeInput).toHaveValue("21");
+  await interfaceSizeInput.fill("12");
+  await interfaceSizeInput.press("Tab");
 
-  await expect(fontSizeInput).toHaveValue("12");
-  await expect(sectionTitle).toHaveCSS("font-size", "9px");
+  await expect(interfaceSizeInput).toHaveValue("12");
+  await expect(contentSizeInput).toHaveValue("21");
+  await expect(sectionTitle).toHaveCSS("font-size", "10px");
 });

packages/app/e2e/browser/archive-finished-subagents.spec.ts

@@ -0,0 +1,79 @@
+import { test } from "../support/fixtures";
+import { openAgentRoute } from "../support/helpers/mock-agent";
+import { seedWorkspace, type SeededWorkspace } from "../support/helpers/seed-client";
+import {
+  archiveFinishedSubagents,
+  expectArchiveFinishedInProgress,
+  expectArchiveFinishedRetry,
+  expectManagedSubagentArchived,
+  expectManagedSubagentUnarchived,
+  expectSubagentRowGone,
+  expectSubagentRowVisible,
+  holdManagedSubagentArchiveRequest,
+  openSubagentsTrack,
+  rejectNextManagedSubagentArchiveRequest,
+  seedParentWithSubagent,
+} from "../support/helpers/subagents";
+
+test.describe("Archive finished subagents", () => {
+  let workspace: SeededWorkspace;
+
+  test.beforeAll(async () => {
+    workspace = await seedWorkspace({ repoPrefix: "archive-finished-subagents-" });
+  });
+
+  test.afterAll(async () => {
+    await workspace?.cleanup();
+  });
+
+  test("archives a finished managed child from the track", async ({ page }) => {
+    const agents = await seedParentWithSubagent(workspace, {
+      parentTitle: "Archive parent",
+      childTitle: "Finished child",
+    });
+    await workspace.client.waitForAgentUpsert(
+      agents.child.id,
+      (snapshot) => snapshot.status === "idle",
+    );
+
+    const archiveGate = await holdManagedSubagentArchiveRequest(page, agents.child.id);
+    await openAgentRoute(page, { workspaceId: agents.workspaceId, agentId: agents.parent.id });
+    await openSubagentsTrack(page);
+    await expectSubagentRowVisible(page, agents.child.id);
+
+    await archiveFinishedSubagents(page);
+    await archiveGate.waitForRequest();
+
+    await expectSubagentRowGone(page, agents.child.id);
+    await expectArchiveFinishedInProgress(page, 0, 1);
+    await expectManagedSubagentUnarchived(workspace, agents.child.id);
+    archiveGate.release();
+    await expectManagedSubagentArchived(workspace, agents.child.id);
+  });
+
+  test("restores a failed child and retries the archive", async ({ page }) => {
+    const agents = await seedParentWithSubagent(workspace, {
+      parentTitle: "Retry parent",
+      childTitle: "Retry child",
+    });
+    await workspace.client.waitForAgentUpsert(
+      agents.child.id,
+      (snapshot) => snapshot.status === "idle",
+    );
+
+    const rejection = await rejectNextManagedSubagentArchiveRequest(page, agents.child.id);
+    await openAgentRoute(page, { workspaceId: agents.workspaceId, agentId: agents.parent.id });
+    await openSubagentsTrack(page);
+    await expectSubagentRowVisible(page, agents.child.id);
+
+    await archiveFinishedSubagents(page);
+    await rejection.waitForRejection();
+
+    await expectSubagentRowVisible(page, agents.child.id);
+    await expectArchiveFinishedRetry(page, 1, 1);
+    await archiveFinishedSubagents(page);
+
+    await expectSubagentRowGone(page, agents.child.id);
+    await expectManagedSubagentArchived(workspace, agents.child.id);
+  });
+});

packages/app/e2e/browser/assistant-selection-copy.spec.ts

@@ -118,6 +118,53 @@ async function selectAssistantText(page: Page, text: string): Promise<void> {
   await selectAssistantTextRange(page, text, text);
 }
 
+async function selectAssistantListItemFromMarker(
+  page: Page,
+  listTag: "ol" | "ul",
+  itemText: string,
+  endText: string,
+): Promise<void> {
+  const assistantMessage = page.getByTestId("assistant-message").filter({
+    hasText: "Direct matches:",
+  });
+  const item = assistantMessage
+    .locator(`[data-paseo-markdown-tag="${listTag}"]`)
+    .filter({ hasText: itemText })
+    .locator(':scope > [data-paseo-markdown-tag="li"]')
+    .filter({ hasText: itemText });
+
+  await item.evaluate((element, selectedEndText) => {
+    const marker = element.querySelector(':scope > [data-paseo-markdown-list-marker="true"]');
+    const markerText = marker?.firstChild;
+    if (!(markerText instanceof Text)) {
+      throw new Error("Expected rendered list marker text");
+    }
+
+    const walker = document.createTreeWalker(element, NodeFilter.SHOW_TEXT);
+    let endNode: Text | null = null;
+    let endOffset = -1;
+    while (walker.nextNode()) {
+      const node = walker.currentNode;
+      const offset = node.textContent?.indexOf(selectedEndText) ?? -1;
+      if (node instanceof Text && offset >= 0) {
+        endNode = node;
+        endOffset = offset + selectedEndText.length;
+        break;
+      }
+    }
+    if (!endNode) {
+      throw new Error(`Could not find list item selection end: ${selectedEndText}`);
+    }
+
+    const range = document.createRange();
+    range.setStart(markerText, 0);
+    range.setEnd(endNode, endOffset);
+    const selection = window.getSelection();
+    selection?.removeAllRanges();
+    selection?.addRange(range);
+  }, endText);
+}
+
 async function doubleClickAssistantMarkdownText(
   page: Page,
   tag: "code" | "strong",
@@ -405,6 +452,24 @@ test("copying an assistant selection preserves Markdown structure and links", as
       '<div>- <strong><a href="https://example.com/issues/1">First issue</a></strong>',
     );
 
+    await selectAssistantListItemFromMarker(page, "ul", "First issue", "failure.");
+    await copySelection(page);
+
+    const draggedBulletClipboard = await readRichClipboard(page);
+    expect(draggedBulletClipboard.plainText).toBe(
+      "- **[First issue](https://example.com/issues/1)**: exact `apply_patch` failure.",
+    );
+    expect(draggedBulletClipboard.html).toContain(
+      '<div>- <strong><a href="https://example.com/issues/1">First issue</a></strong>',
+    );
+
+    await selectAssistantListItemFromMarker(page, "ol", "Sixth item", "Sixth item");
+    await copySelection(page);
+
+    const draggedNumberClipboard = await readRichClipboard(page);
+    expect(draggedNumberClipboard.plainText).toBe("6. Sixth item");
+    expect(draggedNumberClipboard.html).toContain("<div>6. Sixth item</div>");
+
     await selectAssistantText(page, "docs");
     await copySelection(page);
 

packages/app/e2e/browser/bottom-sheet-reopen.spec.ts

@@ -10,6 +10,7 @@ async function openMockAgentAtMobileBreakpoint(page: Page) {
     repoPrefix: "bottom-sheet-reopen-",
     title: "Bottom sheet reopen e2e",
     initialPrompt: "Prepare a bottom sheet reopen test agent.",
+    model: "ten-second-stream",
   });
   await openAgentRoute(page, session);
   await expect(page.getByTestId("workspace-tab-switcher-trigger")).toBeVisible({
@@ -22,16 +23,6 @@ async function openMockAgentAtMobileBreakpoint(page: Page) {
   return session;
 }
 
-async function withMobileMockAgent(page: Page, run: () => Promise<void>) {
-  const session = await openMockAgentAtMobileBreakpoint(page);
-
-  try {
-    await run();
-  } finally {
-    await session.cleanup();
-  }
-}
-
 function bottomSheetBackdrop(page: Page) {
   return page.getByRole("button", { name: "Bottom sheet backdrop" }).first();
 }
@@ -90,11 +81,7 @@ async function openTabSwitcher(page: Page) {
 async function openModelSelector(page: Page) {
   await page.getByRole("button", { name: /Select model/ }).click();
   await expectBottomSheetOpen(page);
-  await expect(
-    page.getByLabel("Bottom Sheet", { exact: true }).getByText("Ten second stream", {
-      exact: true,
-    }),
-  ).toBeVisible({ timeout: 10_000 });
+  await expect(page.getByTestId("agent-controls-settings-list")).toBeVisible({ timeout: 10_000 });
 }
 
 async function openAndCloseTabSwitcherTwice(page: Page) {
@@ -112,27 +99,36 @@ async function openAndCloseModelSelectorTwice(page: Page) {
 }
 
 test.describe("mobile bottom sheet reopen", () => {
-  test("tab switcher can open, close, reopen, and close again", async ({ page }) => {
-    await withMobileMockAgent(page, async () => {
-      await openAndCloseTabSwitcherTwice(page);
-    });
-  });
+  test("sheets reopen and model search returns to configuration", async ({ page }) => {
+    const session = await openMockAgentAtMobileBreakpoint(page);
+    try {
+      await test.step("tab switcher opens, closes, and reopens", async () => {
+        await openAndCloseTabSwitcherTwice(page);
+      });
 
-  test("model selector can open, close, reopen, and close again", async ({ page }) => {
-    await withMobileMockAgent(page, async () => {
-      await openAndCloseModelSelectorTwice(page);
-    });
-  });
+      await test.step("model selector opens, closes, and reopens", async () => {
+        await openAndCloseModelSelectorTwice(page);
+      });
 
-  test("model selector closes after model selection", async ({ page }) => {
-    await withMobileMockAgent(page, async () => {
-      await openModelSelector(page);
-      const sheet = page.getByLabel("Bottom Sheet", { exact: true });
+      await test.step("model search returns to configuration", async () => {
+        await openModelSelector(page);
+        const sheet = page.getByTestId("agent-controls-model-sheet");
 
-      await sheet.getByText("Ten second stream", { exact: true }).click();
+        await page.getByTestId("model-search-all-input").click();
+        const model = page.getByRole("button", { name: /^Ten second stream/ });
+        await expect(model).toBeVisible({
+          timeout: 10_000,
+        });
 
-      await expect(sheet).not.toBeVisible({ timeout: 10_000 });
-      await expect(page.getByRole("button", { name: /Ten second stream/ })).toBeVisible();
-    });
+        await model.click();
+
+        await expect(sheet).toBeVisible();
+        await expect(page.getByTestId("agent-controls-settings-list")).toBeVisible();
+        await expect(page.getByTestId("agent-controls-model")).toContainText("Ten second stream");
+        await expect(page.getByTestId("agent-controls-model-browser-sheet")).not.toBeVisible();
+      });
+    } finally {
+      await session.cleanup();
+    }
   });
 });

packages/app/e2e/browser/branch-switcher.spec.ts

@@ -46,38 +46,6 @@ test.describe("Branch switcher", () => {
   // releases stale sessions from the previous spec; one retry stabilizes it.
   test.describe.configure({ retries: 1 });
 
-  test("switches the workspace branch from the git diff panel for an opaque workspace id", async ({
-    page,
-  }) => {
-    test.setTimeout(90_000);
-    const serverId = getServerId();
-    const workspace = await seedWorkspace({
-      repoPrefix: "branch-switch-",
-      repo: { branches: ["main", "dev"] },
-    });
-
-    try {
-      await gotoAppShell(page);
-      await waitForSidebarHydration(page);
-      await switchWorkspaceViaSidebar({ page, serverId, workspaceId: workspace.workspaceId });
-
-      await openChangesPanel(page);
-      await expectWorkspaceBranch(page, "main");
-      await switchBranchFromChangesPanel(page, { from: "main", to: "dev" });
-      await expectWorkspaceBranch(page, "dev");
-
-      await expect
-        .poll(
-          async () =>
-            (await readWorktreeBranchInfo({ worktreePath: workspace.repoPath })).currentBranch,
-          { timeout: 30_000 },
-        )
-        .toBe("dev");
-    } finally {
-      await workspace.cleanup();
-    }
-  });
-
   test("a custom workspace title stays in the header while the diff panel switches the real branch", async ({
     page,
   }) => {

packages/app/e2e/browser/changes-shortcut-placement.spec.ts

@@ -0,0 +1,27 @@
+import { expect, test } from "../support/fixtures";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+
+const CHANGES_SHORTCUT = `${process.platform === "darwin" ? "Meta" : "Control"}+Shift+G`;
+
+test("Changes shortcut reveals the Changes tree in Explorer", async ({ page }) => {
+  const workspace = await seedWorkspace({ repoPrefix: "changes-shortcut-explorer-" });
+
+  try {
+    await page.setViewportSize({ width: 1400, height: 900 });
+    await gotoWorkspace(page, workspace.workspaceId);
+    await waitForWorkspaceTabsVisible(page);
+
+    await page.keyboard.press(CHANGES_SHORTCUT);
+
+    const explorer = page.getByTestId("workspace-explorer-sidebar").filter({ visible: true });
+    await expect(explorer.getByTestId("explorer-sidebar-tab-changes_tree")).toBeVisible({
+      timeout: 30_000,
+    });
+    await expect(explorer.getByTestId("changes-tree-panel")).toBeVisible();
+    await expect(page.getByTestId("workspace-tab-working_diff")).toHaveCount(0);
+  } finally {
+    await workspace.cleanup();
+  }
+});

packages/app/e2e/browser/chat-outline.spec.ts

@@ -4,6 +4,7 @@ import {
   clickChatOutlineRowEdge,
   disableChatOutlineFromAppearance,
   expectActiveChatOutlinePrompt,
+  expectActiveChatOutlinePromptMovedFrom,
   expectChatOutlinePreview,
   expectChatOutlinePrompts,
   expectChatOutlinePromptToRemainBare,
@@ -29,7 +30,6 @@ import {
   openAgentTimeline,
   scrollThroughOlderHistoryPages,
   scrollTimelineToNewestLoadedEdge,
-  scrollTimelineToOldestLoadedEdge,
   seedLongMockAgentTimeline,
   type LongTimelineAgent,
 } from "../support/helpers/timeline-pagination";
@@ -119,6 +119,7 @@ test.describe("desktop chat outline", () => {
     test("previews and jumps to the focused prompt from the keyboard", async ({ page }) => {
       await focusChatOutlinePrompt(page, 1);
       await expectChatOutlinePreview(page, agent.prompts[0]);
+      await expectTimelinePromptNotMounted(page, agent.oldestPrompt);
 
       await pressEnterOnFocusedPrompt(page);
       await expectTimelinePromptLandedBelowTop(page, agent.oldestPrompt);
@@ -129,11 +130,12 @@ test.describe("desktop chat outline", () => {
       await expectOneActiveChatOutlinePrompt(page);
 
       await clickChatOutlineRowEdge(page, 4);
-      await expectTimelinePromptVisible(page, agent.prompts[3]);
+      await expectTimelinePromptLandedBelowTop(page, agent.prompts[3]);
       await expectActiveChatOutlinePrompt(page, 4);
 
-      await scrollTimelineToOldestLoadedEdge(page);
-      await expectActiveChatOutlinePrompt(page, 1);
+      await scrollTimelineToNewestLoadedEdge(page);
+      await expectTimelinePromptVisible(page, agent.newestPrompt);
+      await expectActiveChatOutlinePromptMovedFrom(page, 4);
     });
 
     test("keeps a clicked prompt free of persistent selection chrome", async ({ page }) => {

packages/app/e2e/browser/command-center-agent-controls.spec.ts

@@ -1,8 +1,10 @@
 import { test } from "../support/fixtures";
+import type { FormPreferences } from "@/create-agent-preferences/preferences";
 import { closeCommandCenter, openCommandCenter } from "../support/helpers/command-center";
 import {
   applyCommandCenterAgentControls,
   chooseCommandCenterAgentControl,
+  openCommandCenterForAgent,
   expectCommandCenterAgentControlSelected,
   expectFocusedAgentControls,
   expectWorkspaceAgentConfiguration,
@@ -11,18 +13,18 @@ import {
 } from "../support/helpers/command-center-agent-controls";
 import { clickNewChat, gotoWorkspace } from "../support/helpers/launcher";
 import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
+import { expectAppRoute } from "../support/helpers/route-assertions";
 import { seedWorkspace } from "../support/helpers/seed-client";
-import { getServerId } from "../support/helpers/server-id";
+import { buildSchedulesRoute } from "@/utils/host-routes";
 
 const CREATE_AGENT_PREFERENCES_KEY = "@paseo:create-agent-preferences";
 
 async function seedMockDraftPreferences(page: import("@playwright/test").Page): Promise<void> {
   await page.addInitScript(
-    ({ preferencesKey, serverId }) => {
+    ({ preferencesKey }) => {
       localStorage.setItem(
         preferencesKey,
         JSON.stringify({
-          serverId,
           provider: "mock",
           providerPreferences: {
             mock: {
@@ -30,10 +32,10 @@ async function seedMockDraftPreferences(page: import("@playwright/test").Page):
               mode: "load-test",
             },
           },
-        }),
+        } satisfies FormPreferences),
       );
     },
-    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY, serverId: getServerId() },
+    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY },
   );
 }
 
@@ -77,6 +79,107 @@ test.describe("Command Center agent controls", () => {
     }
   });
 
+  test("activates the best-matching row, not the highest-ranked group", async ({ page }) => {
+    // "Mode › Load Test" matches "load" at offset 0; every "Model › Mock Load Test › …" row
+    // matches at offset 5. Models used to outrank modes by fixed group rank, so Enter picked a
+    // model. Asserting both model and mode catches either row winning.
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "command-center-relevance-",
+      title: "Command Center relevance",
+      model: "five-minute-stream",
+    });
+    try {
+      await workspace.client.setAgentMode(workspace.agentId, "approval-test");
+      const input = await openCommandCenterForAgent(page, workspace);
+      await input.fill("load");
+      await page.keyboard.press("Enter");
+
+      await expectWorkspaceAgentConfiguration(workspace, {
+        id: workspace.agentId,
+        provider: "mock",
+        model: "five-minute-stream",
+        modeId: "load-test",
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
+  test("matches query tokens in any order", async ({ page }) => {
+    // "Mode › Load test" flattens to "mode load test", so the reversed query "test load" is never
+    // a substring of it. Only per-token matching finds this row.
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "command-center-tokenized-",
+      title: "Command Center tokenized search",
+      model: "five-minute-stream",
+    });
+    try {
+      await workspace.client.setAgentMode(workspace.agentId, "legacy-mode");
+      await openAgentRoute(page, {
+        workspaceId: workspace.workspaceId,
+        agentId: workspace.agentId,
+      });
+      await openCommandCenter(page);
+      await chooseCommandCenterAgentControl({
+        page,
+        query: "test load",
+        choice: "Mode › Load test",
+      });
+
+      await expectWorkspaceAgentConfiguration(workspace, {
+        id: workspace.agentId,
+        provider: "mock",
+        model: "five-minute-stream",
+        modeId: "load-test",
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
+  test("keeps an action first when the action is the best match", async ({ page }) => {
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "command-center-action-priority-",
+      title: "Command Center action priority",
+      model: "five-minute-stream",
+    });
+    try {
+      const input = await openCommandCenterForAgent(page, workspace);
+      await input.fill("sched");
+      await page.keyboard.press("Enter");
+
+      await expectAppRoute(page, buildSchedulesRoute(), { timeout: 30_000 });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
+  test("drops an arrow-key selection once the query changes", async ({ page }) => {
+    const workspace = await seedMockAgentWorkspace({
+      repoPrefix: "command-center-active-reset-",
+      title: "Command Center active reset",
+      model: "five-minute-stream",
+    });
+    try {
+      await workspace.client.setAgentMode(workspace.agentId, "approval-test");
+      const input = await openCommandCenterForAgent(page, workspace);
+      await input.fill("loa");
+      // Moves the highlight onto a model row, which still matches after the next keystroke.
+      await page.keyboard.press("ArrowDown");
+      await input.fill("load");
+      await page.keyboard.press("Enter");
+
+      await expectWorkspaceAgentConfiguration(workspace, {
+        id: workspace.agentId,
+        provider: "mock",
+        model: "five-minute-stream",
+        modeId: "load-test",
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
   test("applies draft model and setting choices to the created agent", async ({ page }) => {
     const workspace = await seedWorkspace({ repoPrefix: "command-center-draft-controls-" });
     await seedMockDraftPreferences(page);

packages/app/e2e/browser/command-center-file-search.spec.ts

@@ -0,0 +1,139 @@
+import { expect, test } from "../support/fixtures";
+import { openCommandCenter } from "../support/helpers/command-center";
+import {
+  delayDirectorySuggestionResponses,
+  expectOneLineTruncatedFileResult,
+  expectStableCommandCenterLayout,
+  failDirectorySuggestionRequests,
+  startCommandCenterLayoutObservation,
+} from "../support/helpers/command-center-file-search";
+import { expectFileTabOpen } from "../support/helpers/file-explorer";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace, type SeededWorkspace } from "../support/helpers/seed-client";
+
+const FILE_PATH =
+  "packages/app/src/command-center/features/workspace-file-search/layout-stability/needle-layout-stability-command-center.tsx";
+const FILE_NAME = "needle-layout-stability-command-center.tsx";
+const FILE_DIRECTORY =
+  "packages/app/src/command-center/features/workspace-file-search/layout-stability";
+
+function paseoSizedFiles(): Array<{ path: string; content: string }> {
+  const files = Array.from({ length: 120 }, (_, index) => ({
+    path: `packages/app/src/features/feature-${String(index).padStart(3, "0")}/index.ts`,
+    content: `export const feature${index} = ${index};\n`,
+  }));
+  files.push({ path: FILE_PATH, content: "export const layoutNeedle = true;\n" });
+  return files;
+}
+
+async function seedMatchingWorkspaces(count: number): Promise<SeededWorkspace[]> {
+  return Promise.all(
+    Array.from({ length: count }, (_, index) =>
+      seedWorkspace({
+        repoPrefix: `command-center-file-search-error-${String(index).padStart(2, "0")}-`,
+        title: `Home failure ${String(index).padStart(2, "0")}`,
+      }),
+    ),
+  );
+}
+
+test.use({
+  userAgent:
+    "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 Chrome/145.0 Safari/537.36",
+});
+
+test("workspace file search stays geometrically stable through delayed loading and results", async ({
+  page,
+}) => {
+  test.setTimeout(120_000);
+  const seeded = await seedWorkspace({
+    repoPrefix: "command-center-file-search-",
+    title: "Paseo-shaped file search",
+    repo: { files: paseoSizedFiles() },
+  });
+
+  try {
+    await delayDirectorySuggestionResponses(page, 800);
+    await gotoWorkspace(page, seeded.workspaceId);
+    await page.keyboard.press("Meta+P");
+
+    const panel = page.getByTestId("command-center-panel");
+    await expect(panel).toBeVisible({ timeout: 30_000 });
+    await expect(panel.getByRole("button", { name: "Files" })).toBeVisible();
+    await startCommandCenterLayoutObservation(page);
+
+    await panel.getByTestId("command-center-input").fill(FILE_NAME);
+    const row = panel.getByRole("button", {
+      name: `${FILE_NAME} ${FILE_DIRECTORY}`,
+    });
+    await expect(row).toBeVisible({ timeout: 30_000 });
+    await expectOneLineTruncatedFileResult(row, {
+      filename: FILE_NAME,
+      path: FILE_DIRECTORY,
+    });
+
+    await panel.getByTestId("command-center-input").fill("needle-layout");
+    await page.waitForTimeout(150);
+    expect(await row.isVisible()).toBe(true);
+    await expect(panel.getByTestId("command-center-file-search-loading")).toBeVisible();
+    await expect(panel.getByText("Searching files...", { exact: true })).toHaveCount(0);
+    await expect(panel.getByTestId("command-center-file-search-loading")).toBeHidden();
+    await expectStableCommandCenterLayout(page);
+
+    await row.click();
+    await expectFileTabOpen(page, FILE_PATH);
+  } finally {
+    await seeded.cleanup();
+  }
+});
+
+test("dropping the files scope leaves the search row the same height", async ({ page }) => {
+  const seeded = await seedWorkspace({
+    repoPrefix: "command-center-file-search-scope-",
+    title: "Scope chip height",
+  });
+
+  try {
+    await gotoWorkspace(page, seeded.workspaceId);
+    await page.keyboard.press("Meta+P");
+
+    const panel = page.getByTestId("command-center-panel");
+    await expect(panel).toBeVisible({ timeout: 30_000 });
+    const header = panel.getByTestId("command-center-header");
+    const scoped = await header.boundingBox();
+
+    await page.getByTestId("command-center-files-scope").click();
+    await expect(page.getByTestId("command-center-files-scope")).toHaveCount(0);
+    const unscoped = await header.boundingBox();
+
+    expect(scoped?.height).toBe(unscoped?.height);
+  } finally {
+    await seeded.cleanup();
+  }
+});
+
+test("unscoped command search keeps commands visible and reports file-search failures", async ({
+  page,
+}) => {
+  test.setTimeout(120_000);
+  const seeded = await seedMatchingWorkspaces(9);
+
+  try {
+    await failDirectorySuggestionRequests(page);
+    await gotoWorkspace(page, seeded[0].workspaceId);
+    const panel = await openCommandCenter(page);
+    await panel.getByTestId("command-center-input").fill("home");
+
+    await expect(panel.getByRole("button", { name: "Home", exact: true })).toBeVisible();
+    await expect(panel.getByRole("button", { name: /Home failure 08/ })).toBeVisible();
+    await expect(panel.getByTestId("command-center-file-search-error")).toHaveText(
+      "Error: Test file search transport failure.",
+    );
+    await expect(panel.getByTestId("command-center-file-search-error")).toBeInViewport({
+      ratio: 1,
+    });
+    await expect(panel.getByText("No matches", { exact: true })).toHaveCount(0);
+  } finally {
+    await Promise.allSettled(seeded.map((workspace) => workspace.cleanup()));
+  }
+});

packages/app/e2e/browser/command-center-grouping.spec.ts

@@ -0,0 +1,58 @@
+import { expect, test, type Page } from "../support/fixtures";
+import { gotoAppShell } from "../support/helpers/app";
+import { openCommandCenter } from "../support/helpers/command-center";
+import { seedWorkspace } from "../support/helpers/seed-client";
+
+const GROUP_BY_STATUS = "Group by status";
+const GROUP_BY_PROJECT = "Group by project";
+
+// Result rows carry no testID of their own, so the entries are addressed by their visible label.
+async function runGroupingEntry(page: Page, label: string, absent: string): Promise<void> {
+  const panel = await openCommandCenter(page);
+  await panel.getByTestId("command-center-input").fill("group");
+
+  const entry = panel.getByText(label, { exact: true });
+  await expect(entry).toBeVisible({ timeout: 30_000 });
+  // The entry always names the mode you are not in, so its opposite must never be listed.
+  await expect(panel.getByText(absent, { exact: true })).toHaveCount(0);
+
+  await entry.click();
+  await expect(page.getByTestId("command-center-panel")).not.toBeVisible({ timeout: 30_000 });
+}
+
+test.describe("Command center sidebar grouping", () => {
+  test.describe.configure({ timeout: 120_000 });
+
+  test("flips sidebar grouping and persists the choice across a reload", async ({ page }) => {
+    const seeded = await seedWorkspace({ repoPrefix: "command-center-grouping-" });
+
+    try {
+      await gotoAppShell(page);
+      const projectList = page.getByTestId("sidebar-project-workspace-list-scroll");
+      const statusList = page.getByTestId("sidebar-status-list-scroll");
+      await expect(projectList).toBeVisible({ timeout: 30_000 });
+
+      // Query-only: the grouping entry stays out of the default empty-query list.
+      const panel = await openCommandCenter(page);
+      await expect(panel.getByText(GROUP_BY_STATUS, { exact: true })).toHaveCount(0);
+      await expect(panel.getByText(GROUP_BY_PROJECT, { exact: true })).toHaveCount(0);
+      await page.keyboard.press("Escape");
+      await expect(panel).not.toBeVisible({ timeout: 30_000 });
+
+      await runGroupingEntry(page, GROUP_BY_STATUS, GROUP_BY_PROJECT);
+      await expect(statusList).toBeVisible({ timeout: 30_000 });
+      await expect(projectList).toHaveCount(0);
+
+      // The only assertion that catches a flip which never reached persisted storage.
+      await page.reload();
+      await expect(statusList).toBeVisible({ timeout: 30_000 });
+
+      // Grouping persists, so the run must leave the sidebar back in project mode.
+      await runGroupingEntry(page, GROUP_BY_PROJECT, GROUP_BY_STATUS);
+      await expect(projectList).toBeVisible({ timeout: 30_000 });
+      await expect(statusList).toHaveCount(0);
+    } finally {
+      await seeded.cleanup().catch(() => undefined);
+    }
+  });
+});

packages/app/e2e/browser/command-center-host.spec.ts

@@ -1,53 +1,14 @@
 import { randomUUID } from "node:crypto";
-import { expect } from "@playwright/test";
 import { test } from "../support/fixtures";
 import { gotoAppShell } from "../support/helpers/app";
 import { createIdleAgent } from "../support/helpers/archive-tab";
 import { openCommandCenter } from "../support/helpers/command-center";
-import { addOfflineHostAndReload } from "../support/helpers/hosts";
 import { expectAgentTabActive } from "../support/helpers/launcher";
 import { seedWorkspace } from "../support/helpers/seed-client";
-import { getServerId } from "../support/helpers/server-id";
-
-const PRIMARY_HOST_LABEL = "Primary Host";
-const SECONDARY_HOST_ID = "host-command-center-secondary";
 
 test.describe("Command center host labels", () => {
   test.describe.configure({ timeout: 180_000 });
 
-  test("agent results show the host they live on when more than one host exists", async ({
-    page,
-  }) => {
-    const seeded = await seedWorkspace({ repoPrefix: "command-center-host-" });
-    const title = `cc-host-${randomUUID().slice(0, 8)}`;
-
-    try {
-      const agent = await createIdleAgent(seeded.client, {
-        cwd: seeded.repoPath,
-        workspaceId: seeded.workspaceId,
-        title,
-      });
-
-      // A second (offline) host makes the view multi-host, which is when the host label earns its space.
-      await gotoAppShell(page);
-      await addOfflineHostAndReload(page, {
-        serverId: SECONDARY_HOST_ID,
-        label: "Secondary Host",
-        primaryLabel: PRIMARY_HOST_LABEL,
-      });
-
-      const panel = await openCommandCenter(page);
-
-      // The shared daemon may carry agents from other specs, so target this agent by its id.
-      const row = panel.getByTestId(`command-center-agent-${getServerId()}:${agent.id}`);
-      await expect(row).toBeVisible({ timeout: 30_000 });
-      await expect(row).toContainText(title);
-      await expect(row).toContainText(PRIMARY_HOST_LABEL);
-    } finally {
-      await seeded.cleanup();
-    }
-  });
-
   test("selecting an agent result opens its workspace tab", async ({ page }) => {
     const seeded = await seedWorkspace({ repoPrefix: "command-center-agent-navigation-" });
     const title = `cc-navigation-${randomUUID().slice(0, 8)}`;

packages/app/e2e/browser/commit-diff-panel.spec.ts

@@ -1,7 +1,9 @@
 import { execFileSync } from "node:child_process";
 import { writeFile } from "node:fs/promises";
 import path from "node:path";
+import type { Locator } from "@playwright/test";
 import { test, expect } from "../support/fixtures";
+import { openChangesTreePanel } from "../support/helpers/workspace-tabs";
 
 const COMMIT_SUBJECT = "Show commit timestamps";
 
@@ -13,7 +15,7 @@ test("commit history explains when the workspace has no commits ahead of its bas
   execFileSync("git", ["checkout", "-b", "feature"], { cwd: workspace.repoPath, stdio: "ignore" });
   await workspace.navigateTo();
 
-  await page.getByRole("button", { name: "Open explorer" }).click();
+  await openChangesTreePanel(page);
   const commitsSection = page.getByRole("button", { name: /Commits/i });
   await expect(commitsSection).toBeVisible({ timeout: 30_000 });
   await commitsSection.click();
@@ -33,7 +35,7 @@ test("commit history shows dates and shares diff layout preferences", async ({
   await page.setViewportSize({ width: 1400, height: 900 });
   await workspace.navigateTo();
 
-  await page.getByRole("button", { name: "Open explorer" }).click();
+  await openChangesTreePanel(page);
   const commitsSection = page.getByRole("button", { name: /Commits/i });
   await expect(commitsSection).toBeVisible({ timeout: 30_000 });
   await commitsSection.click();
@@ -49,14 +51,29 @@ test("commit history shows dates and shares diff layout preferences", async ({
   const panel = page.getByTestId("commit-diff-panel").filter({ visible: true });
   await expect(panel.getByTestId("commit-diff-toolbar")).toBeVisible({ timeout: 30_000 });
   const layoutToggle = panel.getByTestId("commit-diff-toggle-layout");
+  const [commitToolbarBox, layoutToggleBox, layoutToggleGlyphBox] = await Promise.all([
+    panel.getByTestId("commit-diff-header").boundingBox(),
+    layoutToggle.boundingBox(),
+    layoutToggle.locator("svg").boundingBox(),
+  ]);
+  if (!commitToolbarBox || !layoutToggleBox || !layoutToggleGlyphBox) {
+    throw new Error("Commit-diff toolbar geometry could not be measured");
+  }
+  expect(commitToolbarBox.height).toBe(36);
+  expect(layoutToggleBox.width).toBe(20);
+  expect(layoutToggleBox.height).toBe(20);
+  expect(layoutToggleGlyphBox.width).toBe(14);
+  expect(layoutToggleGlyphBox.height).toBe(14);
   await expect(layoutToggle).toHaveAccessibleName("Switch to side-by-side diff");
-  await expect(panel.getByTestId("diff-code-row-0")).toBeVisible({ timeout: 30_000 });
+  await expect(panel.getByTestId("git-diff-canvas")).toBeVisible({ timeout: 30_000 });
+  await expectCommitDiffHeaderGeometry(panel);
 
   await layoutToggle.click();
   await expect(layoutToggle).toHaveAccessibleName("Switch to unified diff");
-  await expect(panel.getByTestId("diff-code-row-0")).toHaveCount(0);
+  await expect(panel.locator('[data-testid^="diff-code-row-"]')).toHaveCount(0);
   await expect(panel.getByTestId("diff-file-0-body")).toBeVisible();
 
+  await page.getByTestId(/^workspace-tab-commit_diff_/).hover();
   await page.getByTestId(/^workspace-commit-diff-close-/).click();
   await expect(panel).toHaveCount(0);
   await commitRow.click();
@@ -67,7 +84,7 @@ test("commit history shows dates and shares diff layout preferences", async ({
 
   await page.setViewportSize({ width: 480, height: 900 });
   await expect(panel.getByTestId("commit-diff-toolbar")).toHaveCount(0);
-  await expect(panel.getByTestId("diff-code-row-0")).toBeVisible();
+  await expect(panel.getByTestId("git-diff-canvas")).toBeVisible();
 });
 
 async function createFeatureCommit(repoPath: string): Promise<void> {
@@ -84,3 +101,25 @@ async function createFeatureCommit(repoPath: string): Promise<void> {
     },
   });
 }
+
+async function expectCommitDiffHeaderGeometry(panel: Locator): Promise<void> {
+  const [header, content, name, stat] = await Promise.all([
+    panel.getByTestId("diff-file-0").boundingBox(),
+    panel.getByTestId("diff-file-0-header-content").boundingBox(),
+    panel.getByTestId("diff-file-0-name").boundingBox(),
+    panel.getByTestId("diff-file-0-stat").boundingBox(),
+  ]);
+  expect(header).not.toBeNull();
+  expect(content).not.toBeNull();
+  expect(name).not.toBeNull();
+  expect(stat).not.toBeNull();
+  expect(header!.height).toBeCloseTo(30, 0);
+  expect(content!.x).toBeCloseTo(header!.x, 0);
+  expect(content!.width).toBeCloseTo(header!.width, 0);
+  expect(name!.x - header!.x).toBeCloseTo(12, 0);
+  expect(name!.x + name!.width).toBeLessThan(stat!.x);
+  expect(Math.abs(name!.y + name!.height / 2 - (stat!.y + stat!.height / 2))).toBeLessThanOrEqual(
+    1,
+  );
+  expect(stat!.x + stat!.width).toBeLessThan(header!.x + header!.width);
+}

packages/app/e2e/browser/composer-attachments.spec.ts

@@ -37,6 +37,7 @@ import {
 import { seedWorkspace } from "../support/helpers/seed-client";
 import { hasGithubAuth, createTempGithubRepo } from "../support/helpers/github-fixtures";
 import { getServerId } from "../support/helpers/server-id";
+import { openFileExplorer } from "../support/helpers/file-explorer";
 
 const MINIMAL_PNG = Buffer.from(
   "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==",
@@ -294,6 +295,28 @@ test.describe("Composer attachments", () => {
     }
   });
 
+  test("Escape cancels file creation without interrupting the running agent", async ({ page }) => {
+    test.setTimeout(120_000);
+    const agent = await startRunningMockAgent(page, {
+      prefix: "file-create-escape-",
+      model: "one-minute-stream",
+      prompt: "Stay running while a file draft is cancelled.",
+    });
+    try {
+      await openFileExplorer(page);
+      await page.getByTestId("files-new-file").click();
+      const nameInput = page.getByTestId("file-explorer-name-input");
+      await expect(nameInput).toBeVisible();
+
+      await nameInput.press("Escape");
+
+      await expect(nameInput).toBeHidden();
+      await expect(page.getByRole("button", { name: /stop|cancel/i }).first()).toBeVisible();
+    } finally {
+      await agent.cleanup();
+    }
+  });
+
   test("composer is locked while new workspace agent is being created", async ({ page }) => {
     test.setTimeout(120_000);
     const serverId = getServerId();

packages/app/e2e/browser/composer-autocomplete.spec.ts

@@ -544,6 +544,88 @@ test.describe("Composer autocomplete", () => {
     }
   });
 
+  test("uses the live cursor when accepting a slash command", async ({ page }) => {
+    await installListCommandsStub(page);
+    const agent = await openReadyMockAgent(page);
+
+    try {
+      const input = composerLocator(page);
+      await expect(input).toBeEditable({ timeout: 30_000 });
+      await input.fill("/tdd");
+      await expect(
+        page
+          .getByTestId("composer-autocomplete-popover")
+          .getByText("/tdd", { exact: true })
+          .first(),
+      ).toBeVisible({ timeout: 30_000 });
+
+      await input.evaluate((element) => {
+        if (!(element instanceof HTMLTextAreaElement)) {
+          throw new Error("Composer input is not a textarea");
+        }
+        element.setSelectionRange(0, 0);
+        element.dispatchEvent(new Event("select", { bubbles: true }));
+        element.dispatchEvent(
+          new KeyboardEvent("keydown", {
+            key: "Tab",
+            bubbles: true,
+            cancelable: true,
+          }),
+        );
+      });
+
+      await expect(input).toHaveValue("/tdd");
+      const selectionStart = await input.evaluate(
+        (element) => (element as HTMLTextAreaElement).selectionStart,
+      );
+      expect(selectionStart).toBe(0);
+    } finally {
+      await agent.cleanup();
+    }
+  });
+
+  test("uses the live input when clicking a slash command", async ({ page }) => {
+    await installListCommandsStub(page);
+    const agent = await openReadyMockAgent(page);
+
+    try {
+      const input = composerLocator(page);
+      await expect(input).toBeEditable({ timeout: 30_000 });
+      await input.fill("/td");
+      const option = page
+        .getByTestId("composer-autocomplete-popover")
+        .getByText("/tdd", { exact: true })
+        .first();
+      await expect(option).toBeVisible({ timeout: 30_000 });
+
+      await option.evaluate((element) => {
+        const textarea = document.querySelector('textarea[aria-label="Message agent..."]');
+        if (!(textarea instanceof HTMLTextAreaElement)) {
+          throw new Error("Composer input is not a textarea");
+        }
+        const valueSetter = Object.getOwnPropertyDescriptor(
+          HTMLTextAreaElement.prototype,
+          "value",
+        )?.set;
+        if (!valueSetter) throw new Error("Textarea value setter is unavailable");
+        valueSetter.call(textarea, "/tdd suffix");
+        textarea.setSelectionRange(4, 4);
+        textarea.dispatchEvent(
+          new InputEvent("input", {
+            bubbles: true,
+            data: "d suffix",
+            inputType: "insertText",
+          }),
+        );
+        (element as HTMLElement).click();
+      });
+
+      await expect(input).toHaveValue("/tdd suffix");
+    } finally {
+      await agent.cleanup();
+    }
+  });
+
   test("stays anchored to the composer when the desktop sidebar is open", async ({ page }) => {
     await installListCommandsStub(page);
     const agent = await openReadyMockAgent(page);

packages/app/e2e/browser/composer-diff-stat.spec.ts

@@ -0,0 +1,127 @@
+import { rm, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { expect, test, type Page } from "../support/fixtures";
+import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
+import { openFilesPanel } from "../support/helpers/workspace-tabs";
+
+const APP_SETTINGS_KEY = "@paseo:app-settings";
+
+function visibleMainPane(page: Page) {
+  return page.getByTestId("workspace-pane-main").filter({ visible: true });
+}
+
+async function seedChangedAgent(repoPrefix: string) {
+  const workspace = await seedMockAgentWorkspace({
+    repoPrefix,
+    title: "Composer diff stat",
+    repo: { withRemote: true },
+  });
+  try {
+    await rm(path.join(workspace.cwd, "remote.git"), { recursive: true });
+    await writeFile(
+      path.join(workspace.cwd, "README.md"),
+      "# Temp Repo\nexport const one = 1;\nexport const two = 2;\n",
+    );
+    await workspace.client.checkoutRefresh(workspace.cwd);
+    await expect
+      .poll(async () => {
+        const workspaces = await workspace.client.fetchWorkspaces();
+        return (
+          workspaces.entries.find((entry) => entry.id === workspace.workspaceId)?.diffStat ?? null
+        );
+      })
+      .toEqual({ additions: 2, deletions: 0 });
+    return workspace;
+  } catch (error) {
+    await workspace.cleanup();
+    throw error;
+  }
+}
+
+test("composer diff stat opens Changes in the configured side pane", async ({ page }) => {
+  await page.addInitScript((settingsKey) => {
+    localStorage.setItem(settingsKey, JSON.stringify({ openInSidePane: { changesLinks: true } }));
+  }, APP_SETTINGS_KEY);
+  const workspace = await seedChangedAgent("composer-diff-stat-side-");
+
+  try {
+    await page.setViewportSize({ width: 1400, height: 900 });
+    await openAgentRoute(page, {
+      workspaceId: workspace.workspaceId,
+      agentId: workspace.agentId,
+    });
+
+    const pill = page.getByTestId("composer-diff-stat-pill");
+    await expect(pill).toBeVisible({ timeout: 30_000 });
+    await expect(pill).toContainText("+2");
+    await expect(pill).toContainText("-0");
+    await pill.click();
+
+    const sidePane = page
+      .locator('[data-testid^="workspace-pane-"]')
+      .filter({ visible: true })
+      .filter({ has: page.getByTestId("working-diff-panel") });
+    await expect(sidePane.getByTestId("workspace-tab-working_diff")).toBeVisible({
+      timeout: 30_000,
+    });
+    await expect(sidePane.getByTestId("working-diff-panel")).toBeVisible({ timeout: 30_000 });
+    await expect(visibleMainPane(page).getByTestId("working-diff-panel")).toHaveCount(0);
+
+    await test.step("Explorer navigation does not replace the side pane", async () => {
+      await openFilesPanel(page);
+      await expect(page.getByTestId("workspace-explorer-sidebar")).toContainText("Files");
+
+      await pill.click();
+      await expect(sidePane.getByTestId("working-diff-panel")).toBeVisible();
+      await expect(page.getByTestId("workspace-tab-working_diff")).toHaveCount(1);
+    });
+  } finally {
+    await workspace.cleanup();
+  }
+});
+
+test("composer diff stat opens the compact explorer instead of a Changes tab", async ({ page }) => {
+  const workspace = await seedChangedAgent("composer-diff-stat-compact-");
+
+  try {
+    await page.setViewportSize({ width: 390, height: 844 });
+    await openAgentRoute(page, {
+      workspaceId: workspace.workspaceId,
+      agentId: workspace.agentId,
+    });
+
+    await page.getByTestId("composer-diff-stat-pill").click();
+
+    await expect(page.getByTestId("changes-header").filter({ visible: true }).first()).toBeVisible({
+      timeout: 30_000,
+    });
+    await expect(page.getByTestId("workspace-tab-working_diff")).toHaveCount(0);
+  } finally {
+    await workspace.cleanup();
+  }
+});
+
+test("composer diff stat opens Changes in the focused pane by default", async ({ page }) => {
+  const workspace = await seedChangedAgent("composer-diff-stat-tab-");
+
+  try {
+    await page.setViewportSize({ width: 1400, height: 900 });
+    await openAgentRoute(page, {
+      workspaceId: workspace.workspaceId,
+      agentId: workspace.agentId,
+    });
+
+    await page.getByTestId("composer-diff-stat-pill").click();
+
+    const mainPane = visibleMainPane(page);
+    await expect(mainPane.getByTestId("workspace-tab-working_diff")).toBeVisible({
+      timeout: 30_000,
+    });
+    await expect(mainPane.getByTestId("working-diff-panel")).toBeVisible({ timeout: 30_000 });
+    await expect(
+      page.locator('[data-testid^="workspace-pane-"]').filter({ visible: true }),
+    ).toHaveCount(1);
+  } finally {
+    await workspace.cleanup();
+  }
+});

packages/app/e2e/browser/composer-whitespace.spec.ts

@@ -0,0 +1,122 @@
+import { expect, test, type Page } from "../support/fixtures";
+import {
+  expectNearBottom,
+  scrollAgentChatToBottom,
+  waitForScrollableChat,
+} from "../support/helpers/agent-bottom-anchor";
+import { awaitAssistantMessage } from "../support/helpers/agent-stream";
+import { composerLocator, expectComposerVisible } from "../support/helpers/composer";
+import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
+
+async function composerHeight(page: Page): Promise<number> {
+  return composerLocator(page).evaluate((element) => element.getBoundingClientRect().height);
+}
+
+async function expectComposerHeight(page: Page, expected: number): Promise<void> {
+  await expect.poll(() => composerHeight(page)).toBe(expected);
+}
+
+async function pressComposerKeyAndMeasureNextPaint(page: Page, key: string): Promise<number> {
+  const composer = composerLocator(page);
+  await installNextComposerPaintProbe(page);
+  await composer.press(key);
+  return readNextComposerPaintHeight(page);
+}
+
+async function typeComposerTextAndMeasureNextPaint(page: Page, text: string): Promise<number> {
+  const composer = composerLocator(page);
+  await installNextComposerPaintProbe(page);
+  await composer.pressSequentially(text);
+  return readNextComposerPaintHeight(page);
+}
+
+async function installNextComposerPaintProbe(page: Page): Promise<void> {
+  await composerLocator(page).evaluate((element) => {
+    Reflect.set(
+      globalThis,
+      "__composerNextPaintHeight",
+      new Promise<number>((resolve) => {
+        element.addEventListener(
+          "input",
+          () => {
+            requestAnimationFrame(() => resolve(element.getBoundingClientRect().height));
+          },
+          { once: true },
+        );
+      }),
+    );
+  });
+}
+
+async function readNextComposerPaintHeight(page: Page): Promise<number> {
+  return page.evaluate(() => Reflect.get(globalThis, "__composerNextPaintHeight"));
+}
+
+test("blank composer lines remain present and keep their measured height", async ({ page }) => {
+  const agent = await seedMockAgentWorkspace({
+    repoPrefix: "composer-whitespace-",
+    title: "Composer whitespace",
+  });
+
+  try {
+    await openAgentRoute(page, agent);
+    await expectComposerVisible(page);
+    const composer = composerLocator(page);
+    const blankLines = "\n\n\n\n\n";
+    const collapsedHeight = await composerHeight(page);
+
+    await test.step("grow the composer with blank lines followed by text", async () => {
+      await composer.fill(`${blankLines}x`);
+      await expect(composer).toHaveValue(`${blankLines}x`);
+      await expect.poll(() => composerHeight(page)).toBeGreaterThan(collapsedHeight);
+    });
+
+    await test.step("delete only the text without losing the blank lines or height", async () => {
+      const expandedHeight = await composerHeight(page);
+      await composer.press("Backspace");
+      await expect(composer).toHaveValue(blankLines);
+      await expectComposerHeight(page, expandedHeight);
+    });
+
+    await test.step("grow on the first paint of a newline and stay stable when it receives a glyph", async () => {
+      await composer.fill("alpha\nbeta");
+      const filledLineHeight = await composerHeight(page);
+      const trailingLineHeight = await pressComposerKeyAndMeasureNextPaint(page, "Shift+Enter");
+      expect(trailingLineHeight).toBeGreaterThan(filledLineHeight);
+
+      const filledTrailingLineHeight = await typeComposerTextAndMeasureNextPaint(page, "x");
+      await expect(composer).toHaveValue("alpha\nbeta\nx");
+      expect(filledTrailingLineHeight).toBe(trailingLineHeight);
+    });
+  } finally {
+    await agent.cleanup();
+  }
+});
+
+test("composer growth keeps a bottom-pinned chat at the bottom", async ({ page }) => {
+  test.setTimeout(90_000);
+  const agent = await seedMockAgentWorkspace({
+    repoPrefix: "composer-bottom-anchor-",
+    title: "Composer bottom anchor",
+    initialPrompt: "Produce enough content to make the chat scrollable.",
+    model: "ten-second-stream",
+  });
+
+  try {
+    await openAgentRoute(page, agent);
+    await expectComposerVisible(page);
+    await awaitAssistantMessage(page);
+    await waitForScrollableChat(page, { minScrollableDistance: 200, timeout: 30_000 });
+    await scrollAgentChatToBottom(page);
+
+    const composer = composerLocator(page);
+    await composer.fill("alpha");
+    for (let line = 0; line < 8; line += 1) {
+      await composer.press("Shift+Enter");
+    }
+
+    await expectNearBottom(page);
+  } finally {
+    await agent.cleanup();
+  }
+});

packages/app/e2e/browser/diff-performance-baseline.ts

@@ -0,0 +1,5 @@
+/** Successful Canvas2D POC run at 6× Chromium CPU slowdown on the checked-in fixture. */
+export const DIFF_POC_BASELINE = {
+  medianFrameMs: 25,
+  p95FrameMs: 50,
+} as const;

packages/app/e2e/browser/diff-performance.spec.ts

@@ -0,0 +1,587 @@
+import { writeFile } from "node:fs/promises";
+import path from "node:path";
+import type { CDPSession, Locator, Page, TestInfo } from "@playwright/test";
+import { buildHostWorkspaceRoute } from "../../src/utils/host-routes";
+import { test, expect } from "../support/fixtures";
+import { getServerId } from "../support/helpers/server-id";
+import { connectSeedClient } from "../support/helpers/seed-client";
+import { createTempGitRepo } from "../support/helpers/workspace";
+import { openChangesPanel, waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+import { DIFF_POC_BASELINE } from "./diff-performance-baseline";
+
+const RUN_DIFF_PERF = process.env.PASEO_DIFF_PERF_E2E === "1";
+const diffPerfDescribe = RUN_DIFF_PERF ? test.describe : test.describe.skip;
+const LINE_COUNT_PER_FILE = 1_200;
+const FILE_PATHS = ["src/large-a.ts", "src/large-b.ts"] as const;
+const CPU_SLOWDOWN = Number(process.env.PASEO_DIFF_PERF_CPU_SLOWDOWN ?? 6);
+const CHANGES_PREFERENCES_KEY = "@paseo:changes-preferences";
+const POC_MEDIAN_FRAME_MS = Number(
+  process.env.PASEO_DIFF_POC_MEDIAN_FRAME_MS ?? DIFF_POC_BASELINE.medianFrameMs,
+);
+const POC_P95_FRAME_MS = Number(
+  process.env.PASEO_DIFF_POC_P95_FRAME_MS ?? DIFF_POC_BASELINE.p95FrameMs,
+);
+const MAX_MEDIAN_FRAME_MS = POC_MEDIAN_FRAME_MS * 1.1;
+const MAX_P95_FRAME_MS = POC_P95_FRAME_MS * 1.1;
+
+diffPerfDescribe("Diff canvas performance", () => {
+  test("opening a review stays responsive in a large diff", async ({ page }) => {
+    test.setTimeout(120_000);
+    const repo = await createLargeDiffRepository();
+    const client = await connectSeedClient();
+    let cdp: CDPSession | null = null;
+
+    try {
+      const created = await client.createWorkspace({
+        source: { kind: "directory", path: repo.path },
+      });
+      if (!created.workspace) {
+        throw new Error(created.error ?? "Failed to create diff performance workspace");
+      }
+
+      await configureUnwrappedUnifiedDiff(page);
+      await page.setViewportSize({ width: 1400, height: 900 });
+      await page.goto(buildHostWorkspaceRoute(getServerId(), created.workspace.id));
+      await waitForWorkspaceTabsVisible(page);
+      await openChangesPanel(page);
+      await expect(page.getByTestId("git-diff-canvas")).toBeVisible({ timeout: 30_000 });
+
+      cdp = await page.context().newCDPSession(page);
+      await cdp.send("Emulation.setCPUThrottlingRate", { rate: CPU_SLOWDOWN });
+      const elapsedMs = await openFirstReviewEditor(page);
+      expect(elapsedMs).toBeLessThanOrEqual(200 * CPU_SLOWDOWN);
+    } finally {
+      await cdp?.send("Emulation.setCPUThrottlingRate", { rate: 1 }).catch(() => undefined);
+      await cdp?.detach().catch(() => undefined);
+      await client.close().catch(() => undefined);
+      await repo.cleanup().catch(() => undefined);
+    }
+  });
+
+  test("one canvas stays filled through continuous fast scrolling", async ({ page }, testInfo) => {
+    test.setTimeout(180_000);
+    const repo = await createLargeDiffRepository();
+    const client = await connectSeedClient();
+    let cdp: CDPSession | null = null;
+    let traceStarted = false;
+
+    try {
+      const created = await client.createWorkspace({
+        source: { kind: "directory", path: repo.path },
+      });
+      if (!created.workspace) {
+        throw new Error(created.error ?? "Failed to create diff performance workspace");
+      }
+
+      await configureUnwrappedUnifiedDiff(page);
+      await page.setViewportSize({ width: 1400, height: 900 });
+      await page.goto(buildHostWorkspaceRoute(getServerId(), created.workspace.id));
+      await waitForWorkspaceTabsVisible(page);
+      await openChangesPanel(page);
+      await expect(page.getByTestId("git-diff-canvas")).toBeVisible({ timeout: 30_000 });
+
+      await expectSingleCanvasSurface(page);
+      cdp = await page.context().newCDPSession(page);
+      await cdp.send("Emulation.setCPUThrottlingRate", { rate: CPU_SLOWDOWN });
+      await startChromiumTrace(cdp);
+      traceStarted = true;
+      await startRendererInstrumentation(page);
+
+      const scrollReports = [];
+      const heapUsage = [];
+      for (let cycle = 0; cycle < 3; cycle += 1) {
+        scrollReports.push(await continuouslyScrollCanvas(page));
+        heapUsage.push(await collectHeapUsage(cdp));
+      }
+      await expectContinuousFramesMatchStableRender(
+        page,
+        scrollReports.flatMap((report) => report.frameSignatures),
+      );
+      const frameImages = await captureCanvasAtScrollIntervals(page);
+      for (const [frameIndex, frame] of frameImages.entries()) {
+        await writeFile(testInfo.outputPath(`diff-canvas-frame-${frameIndex}.png`), frame.image);
+        await testInfo.attach(`diff-canvas-frame-${frameIndex}`, {
+          body: frame.image,
+          contentType: "image/png",
+        });
+        expect(frame.image.byteLength).toBeGreaterThan(4_000);
+      }
+
+      await expectIndependentHorizontalOffsetsSurviveVerticalScrolling(page);
+      const renderer = await readRendererInstrumentation(page);
+      const measuredCycles = scrollReports.slice(1);
+      expect(Math.max(...measuredCycles.map((cycle) => cycle.medianFrameMs))).toBeLessThanOrEqual(
+        MAX_MEDIAN_FRAME_MS,
+      );
+      expect(Math.max(...measuredCycles.map((cycle) => cycle.p95FrameMs))).toBeLessThanOrEqual(
+        MAX_P95_FRAME_MS,
+      );
+      for (const cycle of scrollReports) {
+        expect(Math.abs(cycle.maximumScrollTop - cycle.endScrollTop)).toBeLessThanOrEqual(1);
+      }
+      expect(heapUsage.at(-1)!.usedSize).toBeLessThanOrEqual(
+        heapUsage[0]!.usedSize * 1.1 + 1_000_000,
+      );
+      expect(renderer.addedElements).toBe(0);
+      expect(renderer.removedElements).toBe(0);
+      expect(renderer.reactCommits).toBeLessThanOrEqual(2);
+
+      await reloadDiffPresentation(page, { layout: "unified", wrapLines: true });
+      const wrapped = await measureAdditionalPerformanceScenario(page, cdp, testInfo, "wrapped");
+      await reloadDiffPresentation(page, { layout: "split", wrapLines: false });
+      const split = await measureAdditionalPerformanceScenario(page, cdp, testInfo, "split");
+
+      const report = {
+        cpuSlowdown: CPU_SLOWDOWN,
+        pocBaseline: {
+          medianFrameMs: POC_MEDIAN_FRAME_MS,
+          p95FrameMs: POC_P95_FRAME_MS,
+        },
+        sourceFiles: FILE_PATHS.length,
+        sourceLines: FILE_PATHS.length * LINE_COUNT_PER_FILE,
+        scrollReports,
+        heapUsage,
+        renderer,
+        wrapped,
+        split,
+        canvasPngBytes: frameImages.map((frame) => frame.image.byteLength),
+      };
+      await attachPerformanceReport(testInfo, report);
+      await stopChromiumTrace(cdp, testInfo);
+      traceStarted = false;
+
+      await cdp.send("Emulation.setCPUThrottlingRate", { rate: 1 });
+    } finally {
+      if (cdp && traceStarted) {
+        await stopChromiumTrace(cdp, testInfo).catch(() => undefined);
+      }
+      await cdp?.send("Emulation.setCPUThrottlingRate", { rate: 1 }).catch(() => undefined);
+      await cdp?.detach().catch(() => undefined);
+      await client.close().catch(() => undefined);
+      await repo.cleanup().catch(() => undefined);
+    }
+  });
+});
+
+async function measureAdditionalPerformanceScenario(
+  page: Page,
+  cdp: CDPSession,
+  testInfo: TestInfo,
+  name: "wrapped" | "split",
+) {
+  await expectSingleCanvasSurface(page);
+  await startRendererInstrumentation(page);
+  const scrollReports = [];
+  const heapUsage = [];
+  for (let cycle = 0; cycle < 3; cycle += 1) {
+    scrollReports.push(await continuouslyScrollCanvas(page));
+    heapUsage.push(await collectHeapUsage(cdp));
+  }
+  await expectContinuousFramesMatchStableRender(
+    page,
+    scrollReports.flatMap((report) => report.frameSignatures),
+  );
+  const frames = await captureCanvasAtScrollIntervals(page);
+  for (const [index, frame] of frames.entries()) {
+    await writeFile(testInfo.outputPath(`diff-canvas-${name}-frame-${index}.png`), frame.image);
+    await testInfo.attach(`diff-canvas-${name}-frame-${index}`, {
+      body: frame.image,
+      contentType: "image/png",
+    });
+    expect(frame.image.byteLength).toBeGreaterThan(4_000);
+  }
+  const measuredCycles = scrollReports.slice(1);
+  expect(Math.max(...measuredCycles.map((cycle) => cycle.medianFrameMs))).toBeLessThanOrEqual(
+    MAX_MEDIAN_FRAME_MS,
+  );
+  expect(Math.max(...measuredCycles.map((cycle) => cycle.p95FrameMs))).toBeLessThanOrEqual(
+    MAX_P95_FRAME_MS,
+  );
+  for (const cycle of scrollReports) {
+    expect(Math.abs(cycle.maximumScrollTop - cycle.endScrollTop)).toBeLessThanOrEqual(1);
+  }
+  expect(heapUsage.at(-1)!.usedSize).toBeLessThanOrEqual(heapUsage[0]!.usedSize * 1.1 + 1_000_000);
+  const renderer = await readRendererInstrumentation(page);
+  expect(renderer.addedElements).toBe(0);
+  expect(renderer.removedElements).toBe(0);
+  expect(renderer.reactCommits).toBeLessThanOrEqual(2);
+  return {
+    scrollReports,
+    heapUsage,
+    renderer,
+    canvasPngBytes: frames.map((frame) => frame.image.byteLength),
+  };
+}
+
+async function reloadDiffPresentation(
+  page: Page,
+  requestedPresentation: { layout: "unified" | "split"; wrapLines: boolean },
+): Promise<void> {
+  await page.evaluate(
+    ({ preferencesKey, presentation }) => {
+      const current = JSON.parse(localStorage.getItem(preferencesKey) ?? "{}") as object;
+      localStorage.setItem(preferencesKey, JSON.stringify({ ...current, ...presentation }));
+    },
+    { preferencesKey: CHANGES_PREFERENCES_KEY, presentation: requestedPresentation },
+  );
+  await page.reload();
+  await waitForWorkspaceTabsVisible(page);
+  await openChangesPanel(page);
+  await expect(page.getByTestId("git-diff-canvas")).toBeVisible({ timeout: 30_000 });
+}
+
+async function createLargeDiffRepository() {
+  const repo = await createTempGitRepo("diff-perf-", {
+    files: FILE_PATHS.map((filePath, fileIndex) => ({
+      path: filePath,
+      content: `export const baseline_${fileIndex} = 0;\n`,
+    })),
+  });
+  for (const [fileIndex, filePath] of FILE_PATHS.entries()) {
+    const generatedLines = Array.from(
+      { length: LINE_COUNT_PER_FILE },
+      (_, lineIndex) =>
+        `export const generated_${fileIndex}_${lineIndex} = "${lineIndex}-${"wide".repeat(16)}";`,
+    ).join("\n");
+    await writeFile(
+      path.join(repo.path, filePath),
+      `export const baseline_${fileIndex} = 1;\n${generatedLines}\n`,
+    );
+  }
+  return repo;
+}
+
+async function configureUnwrappedUnifiedDiff(page: Page): Promise<void> {
+  await page.addInitScript(
+    ({ preferencesKey }) => {
+      const scope = globalThis as typeof globalThis & {
+        __PASEO_DIFF_REACT_STATS__?: { commits: number };
+      };
+      scope.__PASEO_DIFF_REACT_STATS__ = { commits: 0 };
+      if (!localStorage.getItem(preferencesKey)) {
+        localStorage.setItem(
+          preferencesKey,
+          JSON.stringify({
+            layout: "unified",
+            viewMode: "flat",
+            wrapLines: false,
+            hideWhitespace: false,
+          }),
+        );
+      }
+    },
+    { preferencesKey: CHANGES_PREFERENCES_KEY },
+  );
+}
+
+async function openFirstReviewEditor(page: Page): Promise<number> {
+  const body = page.getByTestId("diff-file-0-body");
+  const canvas = page.getByTestId("git-diff-canvas");
+  const [bodyBounds, fontSize] = await Promise.all([
+    body.boundingBox(),
+    canvas.evaluate((element) => Number.parseFloat(getComputedStyle(element).fontSize)),
+  ]);
+  if (!bodyBounds) throw new Error("Expanded diff body has no bounds");
+  const lineHeight = Math.round(fontSize * 1.5);
+  const startedAt = performance.now();
+  await page.mouse.click(bodyBounds.x + 120, bodyBounds.y + lineHeight * 1.5);
+  await expect(page.getByTestId("inline-review-editor")).toBeVisible();
+  return performance.now() - startedAt;
+}
+
+async function expectSingleCanvasSurface(page: Page): Promise<void> {
+  await expect(page.locator("canvas")).toHaveCount(1);
+  await expect(page.locator('[data-testid^="diff-code-row-"]')).toHaveCount(0);
+  const elementCount = await page
+    .getByTestId("git-diff-canvas-root")
+    .evaluate((root) => root.querySelectorAll("*").length);
+  expect(elementCount).toBeLessThan(150);
+}
+
+async function continuouslyScrollCanvas(page: Page): Promise<{
+  medianFrameMs: number;
+  p95FrameMs: number;
+  maximumFrameMs: number;
+  slowFrames: number;
+  endScrollTop: number;
+  maximumScrollTop: number;
+  frameSignatures: Array<{ scrollTop: number; signature: number }>;
+}> {
+  return page.getByTestId("git-diff-scroll").evaluate(async (element) => {
+    const scroll = element as HTMLElement;
+    const maximumScrollTop = scroll.scrollHeight - scroll.clientHeight;
+    scroll.scrollTop = 0;
+    const frameDurations: number[] = [];
+    const frameSignatures: Array<{ scrollTop: number; signature: number }> = [];
+    let previousFrame = performance.now();
+
+    for (let step = 1; step <= 90; step += 1) {
+      await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+      const now = performance.now();
+      frameDurations.push(now - previousFrame);
+      previousFrame = now;
+      scroll.scrollTop = (maximumScrollTop * step) / 90;
+      scroll.dispatchEvent(new Event("scroll", { bubbles: false }));
+      if (step % 10 === 0) {
+        await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+        const canvas = document.querySelector<HTMLCanvasElement>(
+          '[data-testid="git-diff-canvas"]',
+        )!;
+        frameSignatures.push({
+          scrollTop: scroll.scrollTop,
+          signature: canvasSignature(canvas, scroll),
+        });
+        previousFrame = performance.now();
+      }
+    }
+    await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+
+    const ordered = [...frameDurations].sort((left, right) => left - right);
+    const percentile = (fraction: number) =>
+      ordered[Math.min(ordered.length - 1, Math.floor(ordered.length * fraction))] ?? 0;
+    return {
+      medianFrameMs: percentile(0.5),
+      p95FrameMs: percentile(0.95),
+      maximumFrameMs: Math.max(...frameDurations),
+      slowFrames: frameDurations.filter((duration) => duration > 50).length,
+      endScrollTop: scroll.scrollTop,
+      maximumScrollTop,
+      frameSignatures,
+    };
+
+    function canvasSignature(canvasElement: HTMLCanvasElement, scrollElement: HTMLElement): number {
+      const sample = document.createElement("canvas");
+      sample.width = 64;
+      sample.height = 36;
+      const context = sample.getContext("2d")!;
+      const canvasCssHeight = Number.parseFloat(canvasElement.style.height);
+      const canvasTop = Number.parseFloat(canvasElement.style.top);
+      const pixelRatio = canvasElement.height / canvasCssHeight;
+      const sourceY = (scrollElement.scrollTop - canvasTop) * pixelRatio;
+      context.drawImage(
+        canvasElement,
+        0,
+        sourceY,
+        canvasElement.width,
+        scrollElement.clientHeight * pixelRatio,
+        0,
+        0,
+        sample.width,
+        sample.height,
+      );
+      const pixels = context.getImageData(0, 0, sample.width, sample.height).data;
+      let hash = 2166136261;
+      for (const value of pixels) hash = Math.imul(hash ^ value, 16777619);
+      return hash >>> 0;
+    }
+  });
+}
+
+interface CapturedCanvasFrame {
+  scrollTop: number;
+  image: Buffer;
+}
+
+async function captureCanvasAtScrollIntervals(page: Page): Promise<CapturedCanvasFrame[]> {
+  const scroll = page.getByTestId("git-diff-scroll");
+  const frames: CapturedCanvasFrame[] = [];
+  for (let step = 0; step <= 8; step += 1) {
+    const scrollTop = await scroll.evaluate(async (element, progress) => {
+      const scrollElement = element as HTMLElement;
+      scrollElement.scrollTop =
+        (scrollElement.scrollHeight - scrollElement.clientHeight) * Number(progress);
+      scrollElement.dispatchEvent(new Event("scroll", { bubbles: false }));
+      await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+      return scrollElement.scrollTop;
+    }, step / 8);
+    frames.push({ scrollTop, image: await scroll.screenshot() });
+  }
+  return frames;
+}
+
+async function expectContinuousFramesMatchStableRender(
+  page: Page,
+  samples: Array<{ scrollTop: number; signature: number }>,
+): Promise<void> {
+  const stable = await page.getByTestId("git-diff-scroll").evaluate(async (element, frames) => {
+    const canvas = document.querySelector<HTMLCanvasElement>('[data-testid="git-diff-canvas"]')!;
+    const signatures: number[] = [];
+    for (const frame of frames) {
+      element.scrollTop = frame.scrollTop;
+      element.dispatchEvent(new Event("scroll", { bubbles: false }));
+      await new Promise<void>((resolve) =>
+        requestAnimationFrame(() => requestAnimationFrame(() => resolve())),
+      );
+      const sample = document.createElement("canvas");
+      sample.width = 64;
+      sample.height = 36;
+      const context = sample.getContext("2d")!;
+      drawVisibleCanvas(context, canvas, element as HTMLElement, sample.width, sample.height);
+      const pixels = context.getImageData(0, 0, sample.width, sample.height).data;
+      let hash = 2166136261;
+      for (const value of pixels) hash = Math.imul(hash ^ value, 16777619);
+      signatures.push(hash >>> 0);
+    }
+    return signatures;
+
+    function drawVisibleCanvas(
+      context: CanvasRenderingContext2D,
+      canvasElement: HTMLCanvasElement,
+      scroll: HTMLElement,
+      width: number,
+      height: number,
+    ): void {
+      const canvasCssHeight = Number.parseFloat(canvasElement.style.height);
+      const canvasTop = Number.parseFloat(canvasElement.style.top);
+      const pixelRatio = canvasElement.height / canvasCssHeight;
+      const sourceY = (scroll.scrollTop - canvasTop) * pixelRatio;
+      context.drawImage(
+        canvasElement,
+        0,
+        sourceY,
+        canvasElement.width,
+        scroll.clientHeight * pixelRatio,
+        0,
+        0,
+        width,
+        height,
+      );
+    }
+  }, samples);
+  expect(stable).toEqual(samples.map((sample) => sample.signature));
+}
+
+async function expectIndependentHorizontalOffsetsSurviveVerticalScrolling(
+  page: Page,
+): Promise<void> {
+  const first = page.getByTestId("diff-file-0-horizontal-scroll");
+  const second = page.getByTestId("diff-file-1-horizontal-scroll");
+  const firstOffset = await setHorizontalOffset(first, 240);
+  const secondOffset = await setHorizontalOffset(second, 120);
+  expect(firstOffset).not.toBe(secondOffset);
+  await page.getByTestId("git-diff-scroll").evaluate(async (element) => {
+    element.scrollTop = element.scrollHeight - element.clientHeight;
+    element.dispatchEvent(new Event("scroll", { bubbles: false }));
+    await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+    element.scrollTop = 0;
+    element.dispatchEvent(new Event("scroll", { bubbles: false }));
+    await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+  });
+  await expect.poll(() => first.evaluate((element) => element.scrollLeft)).toBe(firstOffset);
+  await expect.poll(() => second.evaluate((element) => element.scrollLeft)).toBe(secondOffset);
+}
+
+async function setHorizontalOffset(locator: Locator, offset: number) {
+  return locator.evaluate(async (element, nextOffset) => {
+    element.scrollLeft = nextOffset;
+    element.dispatchEvent(new Event("scroll", { bubbles: false }));
+    await new Promise<void>((resolve) => requestAnimationFrame(() => resolve()));
+    return element.scrollLeft;
+  }, offset);
+}
+
+async function startRendererInstrumentation(page: Page): Promise<void> {
+  await page.getByTestId("git-diff-canvas-root").evaluate((root) => {
+    const scope = globalThis as typeof globalThis & {
+      __PASEO_DIFF_REACT_STATS__?: { commits: number };
+      __PASEO_DIFF_RENDERER_STATS__?: {
+        initialCommits: number;
+        addedElements: number;
+        removedElements: number;
+      };
+    };
+    const reactStats = scope.__PASEO_DIFF_REACT_STATS__ ?? { commits: 0, unmounts: 0 };
+    const stats = {
+      initialCommits: reactStats.commits,
+      addedElements: 0,
+      removedElements: 0,
+    };
+    scope.__PASEO_DIFF_RENDERER_STATS__ = stats;
+    new MutationObserver((records) => {
+      for (const record of records) {
+        stats.addedElements += Array.from(record.addedNodes).filter(
+          (node) => node.nodeType === Node.ELEMENT_NODE,
+        ).length;
+        stats.removedElements += Array.from(record.removedNodes).filter(
+          (node) => node.nodeType === Node.ELEMENT_NODE,
+        ).length;
+      }
+    }).observe(root, { childList: true, subtree: true });
+  });
+}
+
+async function readRendererInstrumentation(page: Page): Promise<{
+  reactCommits: number;
+  addedElements: number;
+  removedElements: number;
+}> {
+  return page.evaluate(() => {
+    const scope = globalThis as typeof globalThis & {
+      __PASEO_DIFF_REACT_STATS__?: { commits: number };
+      __PASEO_DIFF_RENDERER_STATS__?: {
+        initialCommits: number;
+        addedElements: number;
+        removedElements: number;
+      };
+    };
+    const react = scope.__PASEO_DIFF_REACT_STATS__!;
+    const renderer = scope.__PASEO_DIFF_RENDERER_STATS__!;
+    return {
+      reactCommits: react.commits - renderer.initialCommits,
+      addedElements: renderer.addedElements,
+      removedElements: renderer.removedElements,
+    };
+  });
+}
+
+async function collectHeapUsage(cdp: CDPSession): Promise<{ usedSize: number; totalSize: number }> {
+  await cdp.send("HeapProfiler.collectGarbage");
+  return cdp.send("Runtime.getHeapUsage");
+}
+
+async function startChromiumTrace(cdp: CDPSession): Promise<void> {
+  await cdp.send("Tracing.start", {
+    categories: [
+      "devtools.timeline",
+      "v8.execute",
+      "blink.user_timing",
+      "disabled-by-default-v8.cpu_profiler",
+    ].join(","),
+    transferMode: "ReturnAsStream",
+  });
+}
+
+async function stopChromiumTrace(cdp: CDPSession, testInfo: TestInfo): Promise<void> {
+  const completed = new Promise<string>((resolve, reject) => {
+    cdp.once("Tracing.tracingComplete", ({ stream }) => {
+      if (stream) resolve(stream);
+      else reject(new Error("Chromium trace completed without a stream"));
+    });
+  });
+  await cdp.send("Tracing.end");
+  const stream = await completed;
+  const chunks: string[] = [];
+  let eof = false;
+  while (!eof) {
+    const chunk = await cdp.send("IO.read", { handle: stream });
+    chunks.push(chunk.data);
+    eof = chunk.eof;
+  }
+  await cdp.send("IO.close", { handle: stream });
+  const trace = chunks.join("");
+  await writeFile(testInfo.outputPath("diff-canvas-chromium-trace.json"), trace);
+  await testInfo.attach("diff-canvas-chromium-trace", {
+    body: trace,
+    contentType: "application/json",
+  });
+}
+
+async function attachPerformanceReport(
+  testInfo: TestInfo,
+  report: Record<string, unknown>,
+): Promise<void> {
+  await testInfo.attach("diff-canvas-performance", {
+    body: JSON.stringify(report, null, 2),
+    contentType: "application/json",
+  });
+}

packages/app/e2e/browser/direct-terminal-shortcut-pane-placement.spec.ts

@@ -0,0 +1,44 @@
+import { expect, test } from "../support/fixtures";
+import { runWorkspaceActionFromCommandCenter } from "../support/helpers/command-center-workspace-actions";
+import { gotoWorkspace, pressDirectNewTabShortcut } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+
+test.describe("Direct terminal shortcut pane placement", () => {
+  test("opens a terminal in the focused pane", async ({ page }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "direct-terminal-shortcut-pane-" });
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+
+      await runWorkspaceActionFromCommandCenter(page, "Split pane right");
+      const focusedPaneChild = page
+        .getByTestId("split-group-child")
+        .filter({ has: page.getByTestId("workspace-new-tab-panel") });
+      const originalPaneChild = page
+        .getByTestId("split-group-child")
+        .filter({ hasNot: page.getByTestId("workspace-new-tab-panel") });
+      await expect(focusedPaneChild).toBeVisible();
+      const focusedPaneId = await focusedPaneChild
+        .locator('[data-testid^="workspace-pane-"]')
+        .getAttribute("data-testid");
+      const originalPaneId = await originalPaneChild
+        .locator('[data-testid^="workspace-pane-"]')
+        .getAttribute("data-testid");
+      expect(focusedPaneId).not.toBeNull();
+      expect(originalPaneId).not.toBeNull();
+      const focusedPane = page.getByTestId(focusedPaneId!);
+      const originalPane = page.getByTestId(originalPaneId!);
+
+      await pressDirectNewTabShortcut(page, "t");
+
+      await expect(focusedPane.locator('[data-testid^="workspace-tab-terminal_"]')).toHaveCount(1, {
+        timeout: 30_000,
+      });
+      await expect(originalPane.locator('[data-testid^="workspace-tab-terminal_"]')).toHaveCount(0);
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/empty-pane-keyboard.spec.ts

@@ -0,0 +1,65 @@
+import { expect, type Locator, type Page } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { runWorkspaceActionFromCommandCenter } from "../support/helpers/command-center-workspace-actions";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+
+function visibleNewTabPanel(page: Page): Locator {
+  return page.getByTestId("workspace-new-tab-panel").filter({ visible: true });
+}
+
+async function expectLauncherSelection(launcher: Locator, name: string): Promise<void> {
+  await expect(launcher.getByRole("button", { name, exact: true })).toBeFocused();
+}
+
+async function openExplorerWithKeyboard(page: Page): Promise<Locator> {
+  const modifier = process.platform === "darwin" ? "Meta" : "Control";
+  await page.keyboard.press(`${modifier}+E`);
+  const explorer = page.getByTestId("workspace-explorer-sidebar").filter({ visible: true });
+  await expect(explorer).toBeVisible({ timeout: 30_000 });
+  return explorer;
+}
+
+test.describe("New tab keyboard launcher", () => {
+  test("Cmd+E reveals the repository Changes tree", async ({ page }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "explorer-keyboard-" });
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+
+      const explorer = await openExplorerWithKeyboard(page);
+      await expect(explorer.getByTestId("explorer-sidebar-tab-changes_tree")).toBeVisible();
+      await expect(explorer.getByTestId("changes-tree-panel")).toBeVisible();
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
+  test("a new split pane focuses its launcher and supports arrow navigation", async ({ page }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "empty-pane-split-keyboard-" });
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+      await runWorkspaceActionFromCommandCenter(page, "Split pane right");
+
+      const launcher = visibleNewTabPanel(page);
+      await expect(launcher).toBeVisible({ timeout: 30_000 });
+      await expectLauncherSelection(launcher, "Agent");
+
+      await page.keyboard.press("ArrowDown");
+      await expectLauncherSelection(launcher, "Terminal");
+      await page.keyboard.press("ArrowDown");
+      await expectLauncherSelection(launcher, "Diff");
+      await page.keyboard.press("Enter");
+
+      await expect(
+        page.getByTestId("workspace-tab-working_diff").filter({ visible: true }),
+      ).toBeVisible({ timeout: 30_000 });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/explorer-pane-placement.spec.ts

@@ -0,0 +1,224 @@
+// Locks two behaviours that used to break together on web/desktop:
+// 1) with the explorer pane focused, a newly appearing agent must auto-open into the
+//    MAIN pane in the background — the explorer pane is a background surface and must
+//    never swallow entity tabs or lose the user's focus, and
+// 2) splitting a tab out of a pane leaves a New tab behind, and the app must stay
+//    interactive in the [agent | New | explorer] state.
+import { expect, type Locator, type Page } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import {
+  openChangesTreePanel,
+  waitForWorkspaceTabsVisible,
+} from "../support/helpers/workspace-tabs";
+
+function visible(page: Page, testId: string): Locator {
+  return page.getByTestId(testId).filter({ visible: true });
+}
+
+function agentTabChip(page: Page, agentId: string): Locator {
+  return visible(page, `workspace-tab-agent_${agentId}`);
+}
+
+function draftTabChip(page: Page): Locator {
+  return page.locator('[data-testid^="workspace-tab-draft_"]').filter({ visible: true });
+}
+
+/** The Explorer has its own fixed tab rail, separate from workspace pane rows. */
+function explorerTabRow(page: Page): Locator {
+  return visible(page, "explorer-sidebar-tab-rail");
+}
+
+function mainTabRow(page: Page): Locator {
+  return visible(page, "workspace-tabs-row");
+}
+
+async function selectExplorerChanges(page: Page): Promise<void> {
+  await openChangesTreePanel(page);
+  await expect(visible(page, "changes-tree-panel").first()).toBeVisible();
+}
+
+async function closeSeededDraftInMainPane(page: Page): Promise<void> {
+  const chip = draftTabChip(page).first();
+  await chip.hover();
+  await page
+    .locator('[data-testid^="workspace-draft-close-"]')
+    .filter({ visible: true })
+    .first()
+    .click();
+  await expect(draftTabChip(page)).toHaveCount(0, { timeout: 15_000 });
+}
+
+function newTabPaneChild(page: Page): Locator {
+  return page
+    .getByTestId("split-group-child")
+    .filter({ has: page.getByTestId("workspace-new-tab-panel") })
+    .first();
+}
+
+async function emptyPaneBox(page: Page) {
+  const paneChild = newTabPaneChild(page);
+  await expect(paneChild).toBeVisible({ timeout: 15_000 });
+  const box = await paneChild.boundingBox();
+  expect(box).not.toBeNull();
+  if (!box) throw new Error("New tab pane has no bounding box");
+  return box;
+}
+
+async function paneBoxContainingChip(page: Page, chip: Locator) {
+  const paneChild = page.getByTestId("split-group-child").filter({ has: chip }).first();
+  await expect(paneChild).toBeVisible({ timeout: 15_000 });
+  const box = await paneChild.boundingBox();
+  expect(box).not.toBeNull();
+  if (!box) throw new Error("Pane has no bounding box");
+  return box;
+}
+
+/** dnd-kit pointer drag from a tab chip to an absolute point. */
+async function dragChipTo(
+  page: Page,
+  chip: Locator,
+  target: { x: number; y: number },
+): Promise<void> {
+  const chipBox = await chip.boundingBox();
+  expect(chipBox).not.toBeNull();
+  if (!chipBox) throw new Error("Dragged chip has no bounding box");
+
+  await page.mouse.move(chipBox.x + chipBox.width / 2, chipBox.y + chipBox.height / 2);
+  await page.mouse.down();
+  // Exceed the 8px PointerSensor activation distance before heading to the target.
+  await page.mouse.move(chipBox.x + chipBox.width / 2 + 12, chipBox.y + chipBox.height / 2 + 4);
+  await page.mouse.move(target.x, target.y, { steps: 20 });
+  await page.mouse.up();
+}
+
+test.describe("explorer pane tab placement", () => {
+  test("agents open in the main pane and New-tab splits keep the app usable", async ({
+    page,
+  }, testInfo) => {
+    const consoleErrors: string[] = [];
+    page.on("console", (message) => {
+      if (message.type() === "error") consoleErrors.push(message.text());
+    });
+    page.on("pageerror", (error) => consoleErrors.push(`pageerror: ${String(error)}`));
+
+    const workspace = await seedWorkspace({ repoPrefix: "explorer-pane-placement-" });
+    let agentId = "";
+
+    try {
+      await test.step("empty workspace seeds a draft in the main pane", async () => {
+        await gotoWorkspace(page, workspace.workspaceId);
+        await waitForWorkspaceTabsVisible(page);
+        await expect(draftTabChip(page).first()).toBeVisible({ timeout: 30_000 });
+      });
+
+      await test.step("select Changes in Explorer", async () => {
+        await selectExplorerChanges(page);
+      });
+
+      await test.step("an agent appearing now opens in the main pane, not the explorer pane", async () => {
+        const agent = await workspace.client.createAgent({
+          provider: "mock",
+          cwd: workspace.repoPath,
+          workspaceId: workspace.workspaceId,
+          title: "Explorer Placement Agent",
+          modeId: "load-test",
+          model: "ten-second-stream",
+          // Keep the agent streaming through the drag so pane contents are live,
+          // matching the reporter's "spinners were still going" environment.
+          initialPrompt: "stream please",
+        });
+        agentId = agent.id;
+        await expect(agentTabChip(page, agentId).first()).toBeVisible({ timeout: 30_000 });
+
+        await expect(
+          mainTabRow(page).first().getByTestId(`workspace-tab-agent_${agentId}`),
+        ).toBeVisible();
+        await expect(
+          explorerTabRow(page).first().getByTestId(`workspace-tab-agent_${agentId}`),
+        ).toHaveCount(0);
+        // The background open must not change the selected Explorer view.
+        await expect(visible(page, "changes-tree-panel").first()).toBeVisible();
+      });
+
+      await test.step("close the draft; the agent is the main pane's only tab", async () => {
+        await closeSeededDraftInMainPane(page);
+        await expect(
+          mainTabRow(page).first().locator('[data-testid^="workspace-tab-"]'),
+        ).toHaveCount(1);
+        await testInfo.attach("before-drag", {
+          body: await page.screenshot(),
+          contentType: "image/png",
+        });
+      });
+
+      await test.step("split the agent out of the main pane, leaving it empty", async () => {
+        const chip = agentTabChip(page, agentId).first();
+        const mainBox = await paneBoxContainingChip(page, chip);
+        // Land inside the left 15% edge band of the main pane content.
+        await dragChipTo(page, chip, {
+          x: mainBox.x + mainBox.width * 0.06,
+          y: mainBox.y + mainBox.height * 0.6,
+        });
+        await testInfo.attach("after-split", {
+          body: await page.screenshot(),
+          contentType: "image/png",
+        });
+        // The split must have taken: agent pane + New main. Explorer keeps its own rail.
+        await expect(visible(page, "workspace-tabs-row")).toHaveCount(2, { timeout: 10_000 });
+        await expect(explorerTabRow(page)).toHaveCount(1);
+        await expect(page.getByTestId("workspace-new-tab-panel")).toBeVisible();
+      });
+
+      await test.step("app must stay interactive after the split", async () => {
+        // JS main thread alive?
+        const pong = await page.evaluate("1 + 1");
+        expect(pong).toBe(2);
+
+        // Pointer interaction alive? Selecting the agent chip must still work.
+        const chip = agentTabChip(page, agentId).first();
+        await chip.click({ position: { x: 12, y: 13 }, timeout: 5_000 });
+        await expect(chip).toHaveAttribute("aria-selected", "true", { timeout: 5_000 });
+
+        // Clicking inside the New tab pane (focuses it) must not lock anything. Aim
+        // above the vertically centred launcher, which would open a tab instead.
+        const emptyPane = newTabPaneChild(page);
+        await emptyPane.click({ position: { x: 40, y: 120 }, timeout: 5_000 });
+
+        // A second drag must work: center-drop the agent tab back into the New pane.
+        const target = await emptyPaneBox(page);
+        await dragChipTo(page, chip, {
+          x: target.x + target.width / 2,
+          y: target.y + target.height / 2,
+        });
+        await testInfo.attach("after-center-drop", {
+          body: await page.screenshot(),
+          contentType: "image/png",
+        });
+
+        // The agent chip must have replaced New in that pane and stay clickable.
+        await expect(page.getByTestId("workspace-new-tab-panel")).toHaveCount(0, {
+          timeout: 10_000,
+        });
+        await agentTabChip(page, agentId)
+          .first()
+          .click({ position: { x: 12, y: 13 } });
+
+        // The explorer toggle must still respond.
+        await visible(page, "workspace-explorer-toggle").first().click({ timeout: 5_000 });
+        await testInfo.attach("after-interactions", {
+          body: await page.screenshot(),
+          contentType: "image/png",
+        });
+      });
+
+      await testInfo.attach("console-errors", {
+        body: JSON.stringify(consoleErrors, null, 2),
+        contentType: "application/json",
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/explorer-sidebar.spec.ts

@@ -0,0 +1,48 @@
+import { expect, test } from "../support/fixtures";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import {
+  ensureExplorerSidebar,
+  openFilesPanel,
+  waitForWorkspaceTabsVisible,
+} from "../support/helpers/workspace-tabs";
+
+function explorerSidebar(page: Parameters<typeof ensureExplorerSidebar>[0]) {
+  return page.getByTestId("workspace-explorer-sidebar").filter({ visible: true });
+}
+
+test.describe("Explorer sidebar", () => {
+  test("starts with Files and Changes, switches views, and toggles without changing main", async ({
+    page,
+  }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "explorer-sidebar-defaults-" });
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+      const mainTabsBefore = await page
+        .getByTestId("workspace-pane-main")
+        .locator('[data-testid^="workspace-tab-"]')
+        .count();
+
+      const explorer = await ensureExplorerSidebar(page);
+      await expect(explorer.getByTestId("explorer-sidebar-tab-files")).toBeVisible();
+      await expect(explorer.getByTestId("explorer-sidebar-tab-changes_tree")).toBeVisible();
+      await expect(explorer.getByTestId("workspace-new-tab-button")).toHaveCount(0);
+
+      await openFilesPanel(page);
+      await expect(explorer.getByTestId("file-explorer-tree-scroll")).toBeVisible();
+
+      await explorer.getByTestId("explorer-sidebar-tab-changes_tree").click();
+      await expect(explorer.getByTestId("changes-tree-panel")).toBeVisible();
+
+      await page.getByTestId("workspace-explorer-toggle").first().click();
+      await expect(explorerSidebar(page)).toHaveCount(0);
+      await expect(
+        page.getByTestId("workspace-pane-main").locator('[data-testid^="workspace-tab-"]'),
+      ).toHaveCount(mainTabsBefore);
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/explorer-surface-upgrade.spec.ts

@@ -0,0 +1,125 @@
+// Reported on 0.5.0-beta.1 desktop: after upgrading, the old docked explorer
+// sidebar and the new explorer pane were both visible, and on Windows the
+// sidebar had no close button because the window controls own that corner.
+//
+// The sidebar rendered off a persisted `desktop.fileExplorerOpen` flag that no
+// migration cleared and that the rerouted explorer toggle no longer wrote. The
+// explorer is now one surface per layout, so that flag must not resurrect it.
+import { expect, type Locator, type Page } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { waitForWorkspaceTabsVisible } from "../support/helpers/workspace-tabs";
+
+function visible(page: Page, testId: string): Locator {
+  return page.getByTestId(testId).filter({ visible: true });
+}
+
+function readPersistedPanelState(page: Page): Promise<string | null> {
+  return page.evaluate(() => localStorage.getItem("panel-state"));
+}
+
+function explorerSidebar(page: Page): Locator {
+  return visible(page, "workspace-explorer-sidebar");
+}
+
+/** Seed exactly what a pre-pane build wrote after the user opened the sidebar once. */
+async function seedRetiredSidebarState(page: Page): Promise<void> {
+  await page.addInitScript(() => {
+    localStorage.setItem(
+      "panel-state",
+      JSON.stringify({
+        version: 14,
+        state: {
+          desktop: { agentListOpen: true, fileExplorerOpen: true, focusModeEnabled: false },
+          explorerTab: "changes",
+          explorerTabByCheckout: {},
+          expandedPathsByWorkspace: {},
+          diffCollapsedFoldersByWorkspace: {},
+          collapsedFilePathsByWorkspace: {},
+          sidebarWidth: 340,
+          explorerWidth: 400,
+          explorerSortOption: "name",
+          explorerShowHiddenFiles: true,
+          treeRailWidth: 320,
+        },
+      }),
+    );
+  });
+}
+
+test.describe("explorer surface after upgrading from the docked sidebar", () => {
+  test("a persisted fileExplorerOpen flag no longer renders a second explorer", async ({
+    page,
+  }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "explorer-surface-upgrade-" });
+
+    try {
+      await seedRetiredSidebarState(page);
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+
+      await test.step("the retired flag does not reveal Explorer", async () => {
+        await expect(explorerSidebar(page)).toHaveCount(0);
+        const toggle = visible(page, "workspace-explorer-toggle").first();
+        await toggle.click();
+        await expect(explorerSidebar(page)).toHaveCount(1, { timeout: 15_000 });
+        await expect(explorerSidebar(page).getByTestId("explorer-sidebar-tab-files")).toBeVisible();
+        await expect(
+          explorerSidebar(page).getByTestId("explorer-sidebar-tab-changes_tree"),
+        ).toBeVisible();
+
+        await toggle.click();
+        await expect(explorerSidebar(page)).toHaveCount(0, { timeout: 15_000 });
+      });
+
+      await test.step("the retired keys are migrated out of storage", async () => {
+        const persisted = await readPersistedPanelState(page);
+        expect(persisted).not.toBeNull();
+        expect(persisted).not.toContain("fileExplorerOpen");
+        expect(persisted).not.toContain("explorerWidth");
+        // Everything else the user had must survive the migration.
+        expect(persisted).toContain('"sidebarWidth":340');
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+
+  test("toggle-both-sidebars closes and restores the agent list and the explorer", async ({
+    page,
+  }) => {
+    const workspace = await seedWorkspace({ repoPrefix: "explorer-surface-both-" });
+    const modifier = process.platform === "darwin" ? "Meta" : "Control";
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await waitForWorkspaceTabsVisible(page);
+
+      const agentList = visible(page, "sidebar-sessions");
+      const toggle = visible(page, "workspace-explorer-toggle").first();
+
+      await test.step("open the explorer so both sides are showing", async () => {
+        await toggle.click();
+        await expect(explorerSidebar(page)).toHaveCount(1, { timeout: 15_000 });
+        await expect(explorerSidebar(page).getByTestId("explorer-sidebar-tab-files")).toBeVisible();
+        await expect(agentList).toHaveCount(1);
+      });
+
+      await test.step("the shortcut collapses both", async () => {
+        await page.keyboard.press(`${modifier}+.`);
+        await expect(explorerSidebar(page)).toHaveCount(0, { timeout: 15_000 });
+        await expect(agentList).toHaveCount(0);
+      });
+
+      await test.step("the shortcut restores both", async () => {
+        await page.keyboard.press(`${modifier}+.`);
+        await expect(agentList).toHaveCount(1, { timeout: 15_000 });
+        await expect(explorerSidebar(page)).toHaveCount(1, { timeout: 15_000 });
+        await expect(explorerSidebar(page).getByTestId("explorer-sidebar-tab-files")).toBeVisible();
+      });
+    } finally {
+      await workspace.cleanup();
+    }
+  });
+});

packages/app/e2e/browser/file-editing.spec.ts

@@ -9,6 +9,8 @@ import {
 import { installDaemonWebSocketGate } from "../support/helpers/daemon-websocket-gate";
 import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
 
+const APP_SETTINGS_KEY = "@paseo:app-settings";
+
 const RED_PIXEL = Buffer.from(
   "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mP8/x8AAusB9Y9ZQmcAAAAASUVORK5CYII=",
   "base64",
@@ -38,10 +40,7 @@ function fitsViewportWidth(element: HTMLElement): boolean {
 }
 
 async function replaceEditorText(page: Page, content: string): Promise<void> {
-  const contentElement = editor(page);
-  await contentElement.click();
-  await contentElement.press("Control+A");
-  await contentElement.type(content);
+  await editor(page).fill(content);
 }
 
 async function openWorkspaceFile(page: Page, filename: string): Promise<void> {
@@ -93,6 +92,87 @@ async function seedAgentWithFileLink(input: LinkedFile) {
 }
 
 test.describe("CodeMirror workspace file editing", () => {
+  test("renders a lockfile-sized read-only source with a bounded CodeMirror DOM", async ({
+    page,
+  }) => {
+    const session = await seedMockAgentWorkspace({
+      repoPrefix: "file-source-lockfile-",
+      title: "Large source",
+      initialPrompt: "Generate a title and a git branch name. Return JSON only.",
+    });
+    const lockfile = `${'{"packages":['}${Array.from({ length: 42_000 }, (_, index) => `{"name":"package-${index}","version":"1.0.0"}`).join(",")}]}`;
+    await writeFile(path.join(session.cwd, "package-lock.json"), lockfile, "utf8");
+
+    try {
+      await openAgentRoute(page, session);
+      await openWorkspaceFile(page, "package-lock.json");
+
+      await expect(page.getByTestId("file-source-editor")).toBeVisible();
+      await expect(editor(page)).toContainText('"package-0"');
+      await expect.poll(() => page.locator(".cm-line").count()).toBeLessThan(200);
+    } finally {
+      await session.cleanup();
+    }
+  });
+
+  test("keeps the app interactive around a plain 11 MB source", async ({ page }) => {
+    const session = await seedMockAgentWorkspace({
+      repoPrefix: "file-source-plain-",
+      title: "Plain large source",
+      initialPrompt: "Generate a title and a git branch name. Return JSON only.",
+    });
+    await writeFile(
+      path.join(session.cwd, "plain.txt"),
+      "plain source\n".repeat(1_050_000),
+      "utf8",
+    );
+
+    try {
+      await openAgentRoute(page, session);
+      await openWorkspaceFile(page, "plain.txt");
+      await expect(page.getByTestId("file-source-editor")).toBeVisible();
+      await expect(editor(page)).toContainText("plain source");
+      await expect.poll(() => page.locator(".cm-line").count()).toBeLessThan(200);
+
+      await page.getByTestId(`workspace-tab-agent_${session.agentId}`).first().click();
+      await expect(page.getByTestId("message-input-root")).toBeVisible();
+      await page.getByTestId("workspace-tab-file_plain.txt").first().click();
+      await expect(page.getByTestId("file-source-editor")).toBeVisible();
+    } finally {
+      await session.cleanup();
+    }
+  });
+
+  test("refuses a file above the display budget and keeps its tab recoverable", async ({
+    page,
+  }) => {
+    const session = await seedMockAgentWorkspace({
+      repoPrefix: "file-source-unsupported-",
+      title: "Unsupported large source",
+      initialPrompt: "Generate a title and a git branch name. Return JSON only.",
+    });
+    await writeFile(
+      path.join(session.cwd, "too-large.txt"),
+      Buffer.alloc(51 * 1024 * 1024),
+      "utf8",
+    );
+
+    try {
+      await openAgentRoute(page, session);
+      await openWorkspaceFile(page, "too-large.txt");
+      await expect(page.getByTestId("file-source-too-large")).toContainText(
+        "This file is too large to display",
+      );
+
+      await page.getByTestId(`workspace-tab-agent_${session.agentId}`).first().click();
+      await expect(page.getByTestId("message-input-root")).toBeVisible();
+      await page.getByTestId("workspace-tab-file_too-large.txt").first().click();
+      await expect(page.getByTestId("file-source-too-large")).toBeVisible();
+    } finally {
+      await session.cleanup();
+    }
+  });
+
   test("opens an assistant file link at its referenced line", async ({ page }) => {
     const target = "target.ts:42";
     const session = await seedAgentWithFileLink({
@@ -120,7 +200,7 @@ test.describe("CodeMirror workspace file editing", () => {
 
       const sourceEditor = editor(page);
       await sourceEditor.click();
-      await sourceEditor.press("Control+Home");
+      await sourceEditor.press("ControlOrMeta+Home");
       await expect(page.getByLabel(/^Line 1, column \d+$/)).toBeVisible();
 
       await page
@@ -140,6 +220,12 @@ test.describe("CodeMirror workspace file editing", () => {
   });
 
   test("clicking the editor focuses its pane beside an agent", async ({ page }) => {
+    await page.addInitScript((settingsKey) => {
+      localStorage.setItem(
+        settingsKey,
+        JSON.stringify({ openInSidePane: { explorerFiles: true } }),
+      );
+    }, APP_SETTINGS_KEY);
     const target = "target.ts:42";
     const session = await seedAgentWithFileLink({
       target,
@@ -154,9 +240,8 @@ test.describe("CodeMirror workspace file editing", () => {
       await page.setViewportSize({ width: 1280, height: 900 });
       await openAgentRoute(page, session);
 
-      await page.getByRole("button", { name: "Split pane right" }).first().click();
-      await expect(page.getByTestId("workspace-tabs-row").filter({ visible: true })).toHaveCount(2);
       await openWorkspaceFile(page, "target.ts");
+      await expect(page.getByTestId("workspace-tabs-row").filter({ visible: true })).toHaveCount(2);
 
       await page
         .getByTestId(`workspace-tab-agent_${session.agentId}`)
@@ -245,13 +330,13 @@ test.describe("CodeMirror workspace file editing", () => {
     const initialModeBox = await modeControl.boundingBox();
     expect(initialModeBox).not.toBeNull();
     const initialModeX = initialModeBox!.x;
-    await content.press("Control+End");
+    await content.press("ControlOrMeta+End");
     await expect(page.getByLabel(/Line 12, column \d+/)).toBeVisible();
     const movedModeBox = await modeControl.boundingBox();
     expect(movedModeBox).not.toBeNull();
     expect(movedModeBox!.x).toBe(initialModeX);
 
-    await content.press("Control+a");
+    await content.press("ControlOrMeta+a");
     const selection = editorHost.locator(".cm-selectionBackground").first();
     await expect(selection).toBeVisible();
     await expect(selection).toHaveCSS("background-color", "rgba(255, 255, 255, 0.2)");
@@ -347,14 +432,13 @@ test.describe("CodeMirror workspace file editing", () => {
     await replaceEditorText(page, "const localWins = 5;\n");
     await writeFile(sourcePath, "const diskLoses = 6;\n", "utf8");
     await expect(page.getByTestId("file-conflict-alert")).toBeVisible();
+    await page.getByRole("button", { name: "Overwrite", exact: true }).click();
+    await expect.poll(() => readFile(sourcePath, "utf8")).toBe("const localWins = 5;\n");
     for (const fileName of ["one.ts", "two.ts", "three.ts", "four.ts"]) {
       await openWorkspaceFile(page, fileName);
     }
-    await page.getByTestId("workspace-tab-file_source.ts").filter({ visible: true }).click();
+    await openWorkspaceFile(page, "source.ts");
     await expect(editor(page)).toContainText("const localWins = 5;");
-    await expect(page.getByTestId("file-conflict-alert")).toBeVisible();
-    await page.getByRole("button", { name: "Overwrite", exact: true }).click();
-    await expect.poll(() => readFile(sourcePath, "utf8")).toBe("const localWins = 5;\n");
 
     await replaceEditorText(page, "const discarded = 7;\n");
     await writeFile(sourcePath, "const diskWins = 8;\n", "utf8");
@@ -418,10 +502,7 @@ test.describe("CodeMirror workspace file editing", () => {
       .filter({ visible: true })
       .first()
       .click({ button: "right" });
-    await page
-      .getByTestId("workspace-tab-context-file_draft.ts-close")
-      .filter({ visible: true })
-      .click();
+    await page.getByRole("menuitem", { name: "Close", exact: true }).click();
     expect(closePrompt).toContain("Closing it will discard the draft.");
 
     await expect(page.getByTestId("file-source-editor")).toBeVisible();
@@ -441,20 +522,21 @@ test.describe("CodeMirror workspace file editing", () => {
     await workspace.navigateTo();
     await openWorkspaceFile(page, "notes.md");
 
-    await expect(page.getByText("First heading", { exact: true })).toBeVisible();
+    const visibleFilePane = page.getByTestId("workspace-file-pane").filter({ visible: true });
+    await expect(visibleFilePane.getByText("First heading", { exact: true })).toBeVisible();
     await expect(page.getByRole("button", { name: "Preview", exact: true })).toBeVisible();
     await writeFile(markdownPath, "# Updated heading\n", "utf8");
-    await expect(page.getByText("Updated heading", { exact: true })).toBeVisible();
+    await expect(visibleFilePane.getByText("Updated heading", { exact: true })).toBeVisible();
 
     await selectFileView(page, "Source");
     await expect(page.getByTestId("file-source-editor")).toBeVisible();
     await replaceEditorText(page, "# Saved from source\n");
     await expect.poll(() => readFile(markdownPath, "utf8")).toBe("# Saved from source\n");
     await selectFileView(page, "Preview");
-    await expect(page.getByText("Saved from source", { exact: true })).toBeVisible();
+    await expect(visibleFilePane.getByText("Saved from source", { exact: true })).toBeVisible();
 
     await openWorkspaceFile(page, "pixel.png");
-    const image = page.getByTestId("workspace-file-pane").locator("img");
+    const image = visibleFilePane.locator("img");
     await expect(image).toBeVisible();
     const initialSource = await image.getAttribute("src");
     await writeFile(imagePath, BLUE_PIXEL);

packages/app/e2e/browser/file-explorer-context-actions.spec.ts

@@ -0,0 +1,461 @@
+import { mkdir, readFile, rm, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { expect, type Page } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { openFileExplorer } from "../support/helpers/file-explorer";
+import { openChangesPanel } from "../support/helpers/workspace-tabs";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { daemonWsRoutePattern } from "../support/helpers/daemon-port";
+import { seedWorkspace, type SeededWorkspace } from "../support/helpers/seed-client";
+
+let workspace: SeededWorkspace;
+
+async function failNextDeleteRequest(page: Page): Promise<void> {
+  await page.routeWebSocket(daemonWsRoutePattern(), (browserSocket) => {
+    const serverSocket = browserSocket.connectToServer();
+    browserSocket.onMessage((message) => {
+      if (typeof message === "string") {
+        const envelope = JSON.parse(message) as {
+          message?: { type?: string; cwd?: string; path?: string; requestId?: string };
+        };
+        if (envelope.message?.type === "fs.entry.delete.request") {
+          browserSocket.send(
+            JSON.stringify({
+              type: "session",
+              message: {
+                type: "fs.entry.delete.response",
+                payload: {
+                  cwd: envelope.message.cwd,
+                  path: envelope.message.path,
+                  success: false,
+                  error: "Injected delete failure",
+                  requestId: envelope.message.requestId,
+                },
+              },
+            }),
+          );
+          return;
+        }
+      }
+      serverSocket.send(message);
+    });
+    serverSocket.onMessage((message) => browserSocket.send(message));
+  });
+}
+
+async function hideContextActionCapabilities(page: Page): Promise<void> {
+  await page.routeWebSocket(daemonWsRoutePattern(), (browserSocket) => {
+    const serverSocket = browserSocket.connectToServer();
+    browserSocket.onMessage((message) => serverSocket.send(message));
+    serverSocket.onMessage((message) => {
+      if (typeof message !== "string") {
+        browserSocket.send(message);
+        return;
+      }
+      const envelope = JSON.parse(message) as {
+        message?: {
+          type?: string;
+          payload?: { status?: string; features?: Record<string, unknown> };
+        };
+      };
+      if (
+        envelope.message?.type === "status" &&
+        envelope.message.payload?.status === "server_info"
+      ) {
+        envelope.message.payload.features = {
+          ...envelope.message.payload.features,
+          fsEntryOps: false,
+          fsEntryDuplicate: false,
+          checkoutDiscardChanges: false,
+        };
+      }
+      browserSocket.send(JSON.stringify(envelope));
+    });
+  });
+}
+
+async function disconnectNextContextAction(page: Page): Promise<{ wait: () => Promise<void> }> {
+  let resolveRequest: (() => void) | undefined;
+  const requestSeen = new Promise<void>((resolve) => {
+    resolveRequest = resolve;
+  });
+
+  await page.routeWebSocket(daemonWsRoutePattern(), (browserSocket) => {
+    const serverSocket = browserSocket.connectToServer();
+    browserSocket.onMessage((message) => {
+      if (typeof message === "string") {
+        const envelope = JSON.parse(message) as { message?: { type?: string } };
+        const type = envelope.message?.type;
+        if (type?.startsWith("fs.entry.") || type === "checkout.discard_changes.request") {
+          resolveRequest?.();
+          void browserSocket.close({
+            code: 1008,
+            reason: "Context action connection failed",
+          });
+          return;
+        }
+      }
+      serverSocket.send(message);
+    });
+    serverSocket.onMessage((message) => browserSocket.send(message));
+  });
+
+  return { wait: () => requestSeen };
+}
+
+test.beforeEach(async () => {
+  workspace = await seedWorkspace({ repoPrefix: "file-explorer-context-actions-" });
+});
+
+test.afterEach(async () => {
+  await workspace?.cleanup();
+});
+
+test("file explorer rows share the workspace title opacity treatment", async ({ page }) => {
+  await gotoWorkspace(page, workspace.workspaceId);
+  await openFileExplorer(page);
+
+  const row = page
+    .getByTestId("file-explorer-tree-scroll")
+    .getByTestId(/^file-explorer-row-\d+$/)
+    .first();
+  const name = row.getByTestId(/-name$/);
+  await expect(row).toHaveCSS("opacity", "1");
+  await expect(name).toHaveCSS("opacity", "0.76");
+  await row.hover();
+  await expect(name).toHaveCSS("opacity", "1");
+});
+
+test("creates, renames, copies, and deletes entries through the file explorer", async ({
+  context,
+  page,
+}) => {
+  await context.grantPermissions(["clipboard-read", "clipboard-write"]);
+  await gotoWorkspace(page, workspace.workspaceId);
+  await openFileExplorer(page);
+
+  const tree = page.getByTestId("file-explorer-tree-scroll");
+  const entry = (name: string) => tree.getByText(name, { exact: true }).first();
+  const nameInput = page.getByTestId("file-explorer-name-input");
+
+  await page.getByTestId("files-new-folder").click();
+  await expect(nameInput).toBeVisible();
+  const draftRow = nameInput.locator("xpath=..");
+  const placeholderColor = await nameInput.evaluate(
+    (input) => getComputedStyle(input, "::placeholder").color,
+  );
+  const extraMutedChevronColor = await draftRow
+    .locator("svg")
+    .evaluate((icon) => getComputedStyle(icon).stroke);
+  expect(placeholderColor).toBe(extraMutedChevronColor);
+  await nameInput.press("Tab");
+  await expect(nameInput).toBeHidden();
+
+  await page.getByTestId("files-new-file").click();
+  await nameInput.fill("README.md");
+  await nameInput.press("Enter");
+  await expect(page.getByText('"README.md" already exists', { exact: true })).toBeVisible();
+
+  const emptyArea = page.getByTestId("files-empty-area");
+  await expect(emptyArea).toBeVisible();
+  const emptyAreaBounds = await emptyArea.boundingBox();
+  expect(emptyAreaBounds).not.toBeNull();
+  await emptyArea.click({
+    button: "right",
+    position: { x: emptyAreaBounds!.width / 2, y: emptyAreaBounds!.height - 24 },
+  });
+  await expect(page.getByTestId("files-empty-area-new-file")).toBeVisible();
+  await page.getByTestId("files-empty-area-new-folder").click();
+  await nameInput.fill("folder");
+  await nameInput.press("Enter");
+  await expect(entry("folder")).toBeVisible();
+
+  await entry("folder").click({ button: "right" });
+  await page.getByText("New file", { exact: true }).click();
+  await nameInput.fill("child.txt");
+  await nameInput.press("Enter");
+  await expect(entry("child.txt")).toBeVisible();
+
+  await entry("folder").click({ button: "right" });
+  await page.getByText("New folder", { exact: true }).click();
+  await nameInput.fill("nested");
+  await nameInput.press("Enter");
+  await entry("nested").click({ button: "right" });
+  await page.getByText("New folder", { exact: true }).click();
+  await nameInput.fill("deep");
+  await nameInput.press("Enter");
+  await expect(entry("deep")).toBeVisible();
+
+  await entry("folder").click({ button: "right" });
+  await page.getByText("Collapse folder", { exact: true }).click();
+  await expect(entry("child.txt")).toBeHidden();
+  await expect(entry("nested")).toBeHidden();
+  await expect(entry("deep")).toBeHidden();
+
+  await entry("folder").click();
+  await expect(entry("child.txt")).toBeVisible();
+  await expect(entry("nested")).toBeVisible();
+  await expect(entry("deep")).toBeHidden();
+  await entry("nested").click();
+  await expect(entry("deep")).toBeVisible();
+
+  await page.getByTestId("files-new-file").click();
+  await nameInput.fill("created.txt");
+  await nameInput.press("Enter");
+  await expect(entry("created.txt")).toBeVisible();
+  await expect(entry("created.txt")).toHaveCSS("user-select", "none");
+  await entry("created.txt").dblclick();
+  expect(await page.evaluate(() => window.getSelection()?.toString() ?? "")).toBe("");
+
+  const folderRow = entry("folder").locator(
+    "xpath=ancestor::*[starts-with(@data-testid, 'file-explorer-row-')][1]",
+  );
+  const fileRow = entry("created.txt").locator(
+    "xpath=ancestor::*[starts-with(@data-testid, 'file-explorer-row-')][1]",
+  );
+  await expect(folderRow.locator("svg")).toHaveCount(1);
+  await expect(fileRow.locator("svg")).toHaveCount(1);
+  const gitRow = entry(".git").locator(
+    "xpath=ancestor::*[starts-with(@data-testid, 'file-explorer-row-')][1]",
+  );
+  const collapsedChevronBounds = await gitRow.locator("svg").first().boundingBox();
+  const expandedChevronBounds = await folderRow.locator("svg").first().boundingBox();
+  const collapsedChevronSlotBounds = await gitRow
+    .locator("svg")
+    .first()
+    .locator("xpath=../..")
+    .boundingBox();
+  const expandedChevronSlotBounds = await folderRow
+    .locator("svg")
+    .first()
+    .locator("xpath=../..")
+    .boundingBox();
+  expect(collapsedChevronBounds).not.toBeNull();
+  expect(expandedChevronBounds).not.toBeNull();
+  expect(collapsedChevronSlotBounds).not.toBeNull();
+  expect(expandedChevronSlotBounds).not.toBeNull();
+  const collapsedChevronCenter = collapsedChevronBounds!.x + collapsedChevronBounds!.width / 2;
+  const expandedChevronCenter = expandedChevronBounds!.x + expandedChevronBounds!.width / 2;
+  expect(collapsedChevronCenter).toBeCloseTo(expandedChevronCenter, 0);
+  expect(collapsedChevronCenter).toBeCloseTo(
+    collapsedChevronSlotBounds!.x + collapsedChevronSlotBounds!.width / 2,
+    0,
+  );
+  expect(expandedChevronCenter).toBeCloseTo(
+    expandedChevronSlotBounds!.x + expandedChevronSlotBounds!.width / 2,
+    0,
+  );
+  const folderLabelBounds = await entry("folder").boundingBox();
+  const fileLabelBounds = await entry("created.txt").boundingBox();
+  expect(folderLabelBounds).not.toBeNull();
+  expect(fileLabelBounds).not.toBeNull();
+  expect(fileLabelBounds!.x).toBeCloseTo(folderLabelBounds!.x, 0);
+
+  await folderRow.click();
+  await expect(folderRow).toHaveAttribute("aria-selected", "true");
+  await expect(entry("child.txt")).toBeHidden();
+  await folderRow.click();
+  await expect(folderRow).toHaveAttribute("aria-selected", "true");
+  await expect(entry("child.txt")).toBeVisible();
+
+  await entry("folder").click({ button: "right" });
+  await page.getByText("Rename", { exact: true }).click();
+  await nameInput.fill("renamed-folder");
+  await nameInput.press("Tab");
+  await expect(entry("renamed-folder")).toBeVisible();
+  await expect(entry("folder")).toBeHidden();
+  await expect(entry("child.txt")).toBeVisible();
+  const renamedFolderRow = entry("renamed-folder").locator(
+    "xpath=ancestor::*[starts-with(@data-testid, 'file-explorer-row-')][1]",
+  );
+  await expect(renamedFolderRow).toHaveAttribute("aria-selected", "true");
+
+  await entry("created.txt").click({ button: "right" });
+  await expect(fileRow).toHaveAttribute("aria-selected", "true");
+  await expect(renamedFolderRow).toHaveAttribute("aria-selected", "false");
+  const menu = page.getByTestId(/file-explorer-row-\d+-context-menu$/);
+  await expect(menu).toBeVisible();
+
+  await page.getByText("Copy path", { exact: true }).click({ button: "right" });
+  await expect(menu).toBeVisible();
+
+  const backdrop = page.getByTestId(/file-explorer-row-\d+-context-menu-backdrop$/);
+  await backdrop.click({ button: "right" });
+  await expect(menu).toBeHidden();
+
+  await entry("created.txt").click({ button: "right" });
+  await page.getByText("Rename", { exact: true }).click();
+  await expect(nameInput).toHaveValue("created.txt");
+  await nameInput.press("Escape");
+  await expect(entry("created.txt")).toBeVisible();
+
+  await entry("created.txt").click({ button: "right" });
+  await page.getByText("Rename", { exact: true }).click();
+  await nameInput.fill("README.md");
+  await nameInput.press("Enter");
+  await expect(page.getByText('"README.md" already exists', { exact: true })).toBeVisible();
+  await expect(entry("created.txt")).toBeVisible();
+
+  await entry("created.txt").click({ button: "right" });
+  await page.getByText("Rename", { exact: true }).click();
+  await nameInput.fill("renamed.txt");
+  await nameInput.press("Tab");
+  await expect(entry("renamed.txt")).toBeVisible();
+  await expect(entry("created.txt")).toBeHidden();
+
+  await entry("renamed.txt").click({ button: "right" });
+  await page.getByText("Copy relative path", { exact: true }).click();
+  await expect.poll(() => page.evaluate(() => navigator.clipboard.readText())).toBe("renamed.txt");
+
+  await entry("renamed.txt").click({ button: "right" });
+  const deleteAction = page.getByTestId(/file-explorer-row-\d+-delete$/);
+  const deleteLabelColor = await deleteAction
+    .getByText("Delete", { exact: true })
+    .evaluate((element) => getComputedStyle(element).color);
+  await expect(deleteAction.locator("svg")).toHaveCSS("stroke", deleteLabelColor);
+  const cancelledConfirmation = new Promise<string>((resolve) => {
+    page.once("dialog", async (dialog) => {
+      const message = dialog.message();
+      await dialog.dismiss();
+      resolve(message);
+    });
+  });
+  await page.getByText("Delete", { exact: true }).click();
+  expect(await cancelledConfirmation).toContain("renamed.txt");
+  await expect(entry("renamed.txt")).toBeVisible();
+
+  await entry("renamed.txt").click({ button: "right" });
+  const confirmation = new Promise<string>((resolve) => {
+    page.once("dialog", async (dialog) => {
+      const message = dialog.message();
+      await dialog.accept();
+      resolve(message);
+    });
+  });
+  await page.getByText("Delete", { exact: true }).click();
+  expect(await confirmation).toContain("renamed.txt");
+  await expect(entry("renamed.txt")).toBeHidden();
+
+  await entry("renamed-folder").click({ button: "right" });
+  const folderConfirmation = new Promise<string>((resolve) => {
+    page.once("dialog", async (dialog) => {
+      const message = dialog.message();
+      await dialog.accept();
+      resolve(message);
+    });
+  });
+  await page.getByText("Delete", { exact: true }).click();
+  expect(await folderConfirmation).toContain("renamed-folder");
+  await expect(entry("renamed-folder")).toBeHidden();
+  await expect(entry("child.txt")).toBeHidden();
+});
+
+test("duplicates files and folders with collision-free names", async ({ page }) => {
+  const sourceFilePath = path.join(workspace.repoPath, "duplicate-source.txt");
+  const sourceFolderPath = path.join(workspace.repoPath, "duplicate-folder");
+  await writeFile(sourceFilePath, "duplicate contents", "utf8");
+  await mkdir(sourceFolderPath);
+  await writeFile(path.join(sourceFolderPath, "nested.txt"), "nested contents", "utf8");
+
+  try {
+    await gotoWorkspace(page, workspace.workspaceId);
+    await openFileExplorer(page);
+
+    const tree = page.getByTestId("file-explorer-tree-scroll");
+    const entry = (name: string) => tree.getByText(name, { exact: true }).first();
+
+    await entry("duplicate-source.txt").click({ button: "right" });
+    await page.getByTestId(/file-explorer-row-\d+-duplicate$/).click();
+    await expect(entry("duplicate-source copy.txt")).toBeVisible();
+    await expect
+      .poll(() => readFile(path.join(workspace.repoPath, "duplicate-source copy.txt"), "utf8"))
+      .toBe("duplicate contents");
+
+    await entry("duplicate-source.txt").click({ button: "right" });
+    await page.getByTestId(/file-explorer-row-\d+-duplicate$/).click();
+    await expect(entry("duplicate-source copy 2.txt")).toBeVisible();
+
+    await entry("duplicate-folder").click({ button: "right" });
+    await page.getByTestId(/file-explorer-row-\d+-duplicate$/).click();
+    await expect(entry("duplicate-folder copy")).toBeVisible();
+    await expect
+      .poll(() =>
+        readFile(path.join(workspace.repoPath, "duplicate-folder copy", "nested.txt"), "utf8"),
+      )
+      .toBe("nested contents");
+  } finally {
+    await Promise.all([
+      rm(sourceFilePath, { force: true }),
+      rm(path.join(workspace.repoPath, "duplicate-source copy.txt"), { force: true }),
+      rm(path.join(workspace.repoPath, "duplicate-source copy 2.txt"), { force: true }),
+      rm(sourceFolderPath, { recursive: true, force: true }),
+      rm(path.join(workspace.repoPath, "duplicate-folder copy"), { recursive: true, force: true }),
+    ]);
+  }
+});
+
+test("shows an actionable error when the connection drops during creation", async ({ page }) => {
+  const disconnect = await disconnectNextContextAction(page);
+  await gotoWorkspace(page, workspace.workspaceId);
+  await openFileExplorer(page);
+
+  await page.getByTestId("files-new-file").click();
+  const nameInput = page.getByTestId("file-explorer-name-input");
+  await nameInput.fill("disconnected.txt");
+  await nameInput.press("Enter");
+  await disconnect.wait();
+
+  await expect(page.getByText(/Context action connection failed/i)).toBeVisible({
+    timeout: 30_000,
+  });
+  await expect(
+    page.getByTestId("file-explorer-tree-scroll").getByText("disconnected.txt", { exact: true }),
+  ).toHaveCount(0);
+});
+
+test("keeps an entry visible when deletion fails", async ({ page }) => {
+  await failNextDeleteRequest(page);
+  await gotoWorkspace(page, workspace.workspaceId);
+  await openFileExplorer(page);
+
+  const readme = page
+    .getByTestId("file-explorer-tree-scroll")
+    .getByText("README.md", { exact: true });
+  await readme.click({ button: "right" });
+  const confirmation = new Promise<void>((resolve) => {
+    page.once("dialog", async (dialog) => {
+      await dialog.accept();
+      resolve();
+    });
+  });
+  await page.getByText("Delete", { exact: true }).click();
+  await confirmation;
+
+  await expect(page.getByText("Injected delete failure", { exact: true })).toBeVisible();
+  await expect(readme).toBeVisible();
+});
+
+test("hides unsupported file operations and revert actions", async ({ page }) => {
+  await hideContextActionCapabilities(page);
+  await writeFile(path.join(workspace.repoPath, "README.md"), "# Changed while unsupported\n");
+  await gotoWorkspace(page, workspace.workspaceId);
+  await openFileExplorer(page);
+
+  await expect(page.getByTestId("files-new-file")).toHaveCount(0);
+  await expect(page.getByTestId("files-new-folder")).toHaveCount(0);
+  await page
+    .getByTestId("file-explorer-tree-scroll")
+    .getByText("README.md", { exact: true })
+    .click({ button: "right" });
+  const filesMenu = page.getByTestId(/file-explorer-row-\d+-context-menu$/);
+  await expect(filesMenu.getByText("Rename", { exact: true })).toHaveCount(0);
+  await expect(filesMenu.getByText("Duplicate", { exact: true })).toHaveCount(0);
+  await expect(filesMenu.getByText("Delete", { exact: true })).toHaveCount(0);
+  await page.keyboard.press("Escape");
+
+  await openChangesPanel(page);
+  await expect(page.getByTestId("diff-file-0")).toBeVisible({ timeout: 30_000 });
+  await page.getByTestId("diff-file-0-toggle").click({ button: "right" });
+  await expect(page.getByTestId("diff-file-0-duplicate")).toHaveCount(0);
+  await expect(page.getByTestId("diff-file-0-revert")).toHaveCount(0);
+});

packages/app/e2e/browser/file-transition.spec.ts

@@ -0,0 +1,147 @@
+import { expect } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { delayFileReadResponse } from "../support/helpers/file-read-gate";
+import {
+  expandFolder,
+  openFileExplorer,
+  openFileFromExplorer,
+} from "../support/helpers/file-explorer";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace, type SeededWorkspace } from "../support/helpers/seed-client";
+
+let workspace: SeededWorkspace;
+
+test.beforeAll(async () => {
+  workspace = await seedWorkspace({
+    repoPrefix: "file-transition-",
+    repo: {
+      files: [
+        { path: "src/a.ts", content: "export const sourceA = true;\n" },
+        { path: "src/b.ts", content: "export const sourceB = true;\n" },
+        { path: "docs/b.md", content: "# Markdown B\n" },
+      ],
+    },
+  });
+});
+
+test.afterAll(async () => workspace?.cleanup());
+
+async function openSourceA(page: import("@playwright/test").Page) {
+  await openFileExplorer(page);
+  await expandFolder(page, "src");
+  await openFileFromExplorer(page, "a.ts");
+  await expect(page.getByText("export const sourceA = true;")).toBeVisible({ timeout: 30_000 });
+}
+
+async function assertHeldFileTransition(
+  page: import("@playwright/test").Page,
+  targetPath: string,
+  finalText: string,
+  gate: Awaited<ReturnType<typeof delayFileReadResponse>>,
+) {
+  await page.evaluate(() => {
+    const originalChip = document.querySelector('[data-testid="workspace-tab-file_src/a.ts"]');
+    (
+      window as typeof window & { __fileTransitionOriginalChip?: Element | null }
+    ).__fileTransitionOriginalChip = originalChip;
+  });
+  await page.evaluate((tabTestId) => {
+    const timeline: Array<{ chip: boolean; loading: boolean; unavailable: boolean }> = [];
+    const sample = () => {
+      const next = {
+        chip: Boolean(document.querySelector(`[data-testid="${tabTestId}"]`)),
+        loading: Boolean(document.querySelector('[data-testid="file-preview-loading"]')),
+        unavailable: Boolean(document.querySelector('[data-testid="file-preview-unsupported"]')),
+      };
+      if (JSON.stringify(timeline.at(-1)) !== JSON.stringify(next)) timeline.push(next);
+    };
+    sample();
+    const observer = new MutationObserver(sample);
+    observer.observe(document.body, { childList: true, subtree: true, characterData: true });
+    (
+      window as typeof window & { __fileTransitionTimeline?: typeof timeline }
+    ).__fileTransitionTimeline = timeline;
+    (
+      window as typeof window & { __fileTransitionObserver?: MutationObserver }
+    ).__fileTransitionObserver = observer;
+  }, `workspace-tab-file_${targetPath}`);
+  await openFileFromExplorer(page, targetPath.split("/").pop() as string);
+  await gate.waitUntilHeld();
+
+  const targetChip = page.getByTestId(`workspace-tab-file_${targetPath}`).first();
+  await expect(targetChip).toBeVisible();
+  await expect(targetChip).toContainText(targetPath.split("/").pop() as string);
+  await expect(page.getByTestId("file-preview-loading")).toBeVisible();
+  await expect(page.getByTestId("file-preview-unsupported")).toHaveCount(0);
+  await expect(page.getByText("export const sourceA = true;")).toHaveCount(0);
+
+  const chipIdentity = await page.evaluate((tabTestId) => {
+    const windowWithOriginal = window as typeof window & {
+      __fileTransitionOriginalChip?: Element | null;
+    };
+    const originalChip = windowWithOriginal.__fileTransitionOriginalChip ?? null;
+    const currentChip = document.querySelector(`[data-testid="${tabTestId}"]`);
+    return {
+      originalStillConnected: originalChip?.isConnected ?? false,
+      sameNode: originalChip === currentChip,
+    };
+  }, `workspace-tab-file_${targetPath}`);
+  expect(chipIdentity).toEqual({ originalStillConnected: true, sameNode: true });
+
+  const heldTimeline = await page.evaluate(
+    () =>
+      (
+        window as typeof window & {
+          __fileTransitionTimeline?: Array<{
+            chip: boolean;
+            loading: boolean;
+            unavailable: boolean;
+          }>;
+        }
+      ).__fileTransitionTimeline ?? [],
+  );
+  expect(heldTimeline).toContainEqual(expect.objectContaining({ chip: true, loading: true }));
+  expect(heldTimeline).not.toContainEqual(expect.objectContaining({ unavailable: true }));
+  const descriptorIndex = heldTimeline.findIndex((sample) => sample.chip);
+  expect(heldTimeline.slice(descriptorIndex)).not.toContainEqual(
+    expect.objectContaining({ chip: false }),
+  );
+
+  gate.release();
+  await expect(page.getByText(finalText)).toBeVisible({ timeout: 30_000 });
+  await expect(page.getByTestId("file-preview-loading")).toHaveCount(0);
+  const timeline = await page.evaluate(() => {
+    const windowWithTimeline = window as typeof window & {
+      __fileTransitionTimeline?: Array<{ chip: boolean; loading: boolean; unavailable: boolean }>;
+      __fileTransitionObserver?: MutationObserver;
+    };
+    windowWithTimeline.__fileTransitionObserver?.disconnect();
+    return windowWithTimeline.__fileTransitionTimeline ?? [];
+  });
+  expect(
+    timeline.filter((sample, index) => sample.loading && !timeline[index - 1]?.loading),
+  ).toHaveLength(1);
+}
+
+test.describe("File target transitions", () => {
+  test("keeps the Markdown descriptor and loading state stable through a real delayed read", async ({
+    page,
+  }) => {
+    const gate = await delayFileReadResponse(page, "docs/b.md");
+    await page.reload();
+    await gotoWorkspace(page, workspace.workspaceId);
+    await openSourceA(page);
+    await expandFolder(page, "docs");
+    await assertHeldFileTransition(page, "docs/b.md", "Markdown B", gate);
+  });
+
+  test("keeps the source descriptor and loading state stable through a real delayed read", async ({
+    page,
+  }) => {
+    const gate = await delayFileReadResponse(page, "src/b.ts");
+    await page.reload();
+    await gotoWorkspace(page, workspace.workspaceId);
+    await openSourceA(page);
+    await assertHeldFileTransition(page, "src/b.ts", "export const sourceB = true;", gate);
+  });
+});

packages/app/e2e/browser/launcher-tab.spec.ts

@@ -1,17 +1,22 @@
 import { test, expect } from "../support/fixtures";
+import type { Locator } from "@playwright/test";
 import {
   gotoWorkspace,
   assertNewChatTileVisible,
   assertNewTabMenuTriggerVisible,
   assertSingleNewTabButton,
-  pressNewTabShortcut,
+  openNewTabMenuWithShortcut,
   clickNewChat,
   clickNewTerminal,
   countTabsOfKind,
+  waitForTabBar,
+  pressNewTabShortcut,
+  pressDirectNewTabShortcut,
   getTabTestIds,
   waitForTabWithTitle,
   measureTileTransition,
   sampleTabsDuringTransition,
+  expectTabTitleFits,
   terminalSurfaceLocator,
 } from "../support/helpers/launcher";
 import { expectComposerVisible, composerLocator } from "../support/helpers/composer";
@@ -20,13 +25,47 @@ import {
   setupDeterministicPrompt,
 } from "../support/helpers/terminal-perf";
 import { seedWorkspace, type SeededWorkspace } from "../support/helpers/seed-client";
+import {
+  expectTerminalOutputContains,
+  seedTerminalProfiles,
+  type TerminalProfile,
+} from "../support/helpers/new-workspace-launch";
+import { gotoAppShell } from "../support/helpers/app";
+import { waitForSidebarHydration } from "../support/helpers/workspace-ui";
+import { getServerId } from "../support/helpers/server-id";
 
 // ─── Shared state ──────────────────────────────────────────────────────────
 
 let workspace: SeededWorkspace;
+let secondWorkspaceId: string | null = null;
+
+const EMPTY_PROMPT_PROFILE: TerminalProfile = {
+  id: "e2e-empty-prompt",
+  name: "Empty Prompt",
+  command: "/bin/sh",
+  args: ["-c", 'echo prompt-args: "$#"; sleep 10', "profile-name", "{{{prompt}}}"],
+};
+
+async function tabTestIds(tabs: Locator): Promise<(string | null)[]> {
+  return tabs.evaluateAll((elements) =>
+    elements.map((element) => element.getAttribute("data-testid")),
+  );
+}
+
+function tabIdentityKey(snapshot: Array<{ id: string }>): string {
+  return JSON.stringify(snapshot.map(({ id }) => id));
+}
 
 test.beforeAll(async () => {
   workspace = await seedWorkspace({ repoPrefix: "launcher-e2e-" });
+  const created = await workspace.client.createWorkspace({
+    source: { kind: "directory", path: workspace.repoPath, projectId: workspace.projectId },
+    title: "launcher-e2e-cmd-t-switch-target",
+  });
+  if (!created.workspace) {
+    throw new Error(created.error ?? "Failed to create secondary workspace for cmd+t switch test");
+  }
+  secondWorkspaceId = created.workspace.id;
 });
 
 test.afterAll(async () => {
@@ -38,31 +77,128 @@ test.afterAll(async () => {
 // ═══════════════════════════════════════════════════════════════════════════
 
 test.describe("Tab creation", () => {
-  test("Cmd+T opens a new agent tab with composer", async ({ page }) => {
-    await gotoWorkspace(page, workspace.workspaceId);
+  test("Cmd+T keeps creating a New tab after workspace switches", async ({ page }) => {
+    if (!secondWorkspaceId) {
+      throw new Error("Secondary workspace was not created");
+    }
 
-    await pressNewTabShortcut(page);
+    const serverId = getServerId();
+    const switchWorkspaceRow = async (workspaceId: string) => {
+      const row = page.getByTestId(`sidebar-workspace-row-${serverId}:${workspaceId}`).first();
+      await expect(row).toBeVisible({ timeout: 30_000 });
+      await row.click();
+      // The shortcut must follow the route, so prove the route actually moved first.
+      await expect(page).toHaveURL(new RegExp(`workspace/${workspaceId}(\\b|/|$)`), {
+        timeout: 15_000,
+      });
+      await waitForTabBar(page);
+    };
 
-    await expectComposerVisible(page);
+    await gotoAppShell(page);
+    await waitForSidebarHydration(page);
+
+    const sequence = [workspace.workspaceId, secondWorkspaceId];
+    for (let i = 0; i < 8; i++) {
+      await switchWorkspaceRow(sequence[i % sequence.length]);
+      await pressNewTabShortcut(page);
+      await expect(
+        page.getByTestId("workspace-new-tab-panel").filter({ visible: true }),
+        `New tab did not open after switch ${i}`,
+      ).toBeVisible({
+        timeout: 5_000,
+      });
+    }
   });
 
-  test("opening two new tabs creates two draft tabs", async ({ page }) => {
+  test("retained inactive workspaces cannot own New tab shortcuts", async ({ page }) => {
     await gotoWorkspace(page, workspace.workspaceId);
+    const workspaceUrl = page.url();
+    const modifier = process.platform === "darwin" ? "Meta" : "Control";
+    const newTabCountBefore = await countTabsOfKind(page, "new_tab");
 
+    await page.keyboard.press(`${modifier}+Comma`);
+    await expect(page.getByRole("navigation", { name: "Settings" })).toBeVisible();
+
+    await page.keyboard.press(`${modifier}+t`);
+
+    await page.goto(workspaceUrl);
+    await assertNewTabMenuTriggerVisible(page);
+    await expect.poll(() => countTabsOfKind(page, "new_tab")).toBe(newTabCountBefore);
+  });
+
+  test("Cmd+T creates a New tab without creating an agent", async ({ page }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
     const countBefore = await countTabsOfKind(page, "draft");
 
-    await pressNewTabShortcut(page);
-    await expect
-      .poll(() => countTabsOfKind(page, "draft"), { timeout: 15_000 })
-      .toBe(countBefore + 1);
-    const countAfterFirst = await countTabsOfKind(page, "draft");
+    await openNewTabMenuWithShortcut(page);
+
+    await expect.poll(() => countTabsOfKind(page, "draft")).toBe(countBefore);
+    await expect(
+      page.getByTestId("workspace-new-tab-panel").filter({ visible: true }),
+    ).toBeVisible();
+  });
+
+  test("clicking + opens its menu without creating a New tab", async ({ page }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
+    const countBefore = await countTabsOfKind(page, "new_tab");
+
+    await page.getByTestId("workspace-new-tab-button").filter({ visible: true }).click();
+
+    await expect(
+      page.getByTestId("workspace-new-tab-menu").filter({ visible: true }),
+    ).toBeVisible();
+    await expect.poll(() => countTabsOfKind(page, "new_tab")).toBe(countBefore);
+  });
+
+  test("opening two New tabs creates two independent tab identities", async ({ page }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
+
+    const newTabs = page
+      .locator('[data-testid^="workspace-tab-tab_"]')
+      .filter({ hasText: "New tab" });
+    const countBefore = await newTabs.count();
 
-    // Blur the composer so the second shortcut isn't swallowed by the focused input
-    await page.evaluate(() => (document.activeElement as HTMLElement)?.blur?.());
     await pressNewTabShortcut(page);
-    await expect
-      .poll(() => countTabsOfKind(page, "draft"), { timeout: 15_000 })
-      .toBe(countAfterFirst + 1);
+    await expect(newTabs).toHaveCount(countBefore + 1);
+    const firstIds = await tabTestIds(newTabs);
+
+    await pressNewTabShortcut(page);
+    await expect(newTabs).toHaveCount(countBefore + 2);
+    const ids = await tabTestIds(newTabs);
+    expect(new Set(ids).size).toBe(ids.length);
+    expect(ids.filter((id) => !firstIds.includes(id))).toHaveLength(1);
+  });
+
+  test("New tab exposes shortcuts and supports arrow navigation after refocus", async ({
+    page,
+  }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
+    await openNewTabMenuWithShortcut(page);
+
+    const panel = page.getByTestId("workspace-new-tab-panel").filter({ visible: true });
+    const agent = panel.getByRole("button", { name: /^Agent/ });
+    const terminal = panel.getByRole("button", { name: /^Terminal/ });
+    const diff = panel.getByRole("button", { name: /Diff/ });
+    const shortcutPrefix = process.platform === "darwin" ? /⇧⌘/ : /Ctrl.*Shift/;
+    await expect(agent).toContainText(new RegExp(`${shortcutPrefix.source}.*A`));
+    await expect(terminal).toContainText(new RegExp(`${shortcutPrefix.source}.*T`));
+    await expect(diff).toContainText(new RegExp(`${shortcutPrefix.source}.*G`));
+    await expect(agent).toBeFocused();
+
+    await page.keyboard.press("ArrowDown");
+    await expect(terminal).toBeFocused();
+    await page.keyboard.press("ArrowUp");
+    await expect(agent).toBeFocused();
+
+    await pressDirectNewTabShortcut(page, "e");
+    await expect(page.getByTestId("workspace-explorer-sidebar")).toBeVisible();
+    await expect(
+      page.getByTestId("file-explorer-tree-scroll").filter({ visible: true }),
+    ).toBeVisible();
+
+    await page.locator("body").click({ position: { x: 1, y: 1 } });
+    await panel.click({ position: { x: 20, y: 20 } });
+    await expect(agent).toBeFocused();
   });
 
   test("clicking new agent tab creates a draft tab", async ({ page }) => {
@@ -90,12 +226,61 @@ test.describe("Tab creation", () => {
     expect(terminalTabs.length).toBeGreaterThanOrEqual(1);
   });
 
+  test("launching a profile from the New tab menu drops its empty prompt argument", async ({
+    page,
+  }) => {
+    test.setTimeout(45_000);
+    const profileSeed = await seedTerminalProfiles([EMPTY_PROMPT_PROFILE]);
+
+    try {
+      await gotoWorkspace(page, workspace.workspaceId);
+      await page.getByTestId("workspace-new-tab-button").filter({ visible: true }).click();
+      await page
+        .getByTestId("workspace-new-tab-menu")
+        .filter({ visible: true })
+        .getByRole("menuitem", { name: EMPTY_PROMPT_PROFILE.name })
+        .click();
+
+      await expectTerminalOutputContains(page, "prompt-args: 0");
+    } finally {
+      await profileSeed.restore();
+    }
+  });
+
+  test("terminal profiles are grouped with a settings action", async ({ page }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
+    await page.getByTestId("workspace-new-tab-button").filter({ visible: true }).click();
+
+    const menu = page.getByTestId("workspace-new-tab-menu").filter({ visible: true });
+    await expect(menu.getByText("Terminal profiles", { exact: true })).toBeVisible();
+
+    const editProfiles = menu.getByTestId("workspace-new-tab-menu-edit-terminal-profiles");
+    await expect(editProfiles).toHaveAccessibleName("Edit profiles");
+
+    await editProfiles.click();
+    await expect(page).toHaveURL(/\/settings\/hosts\/[^/]+\/terminals$/);
+  });
+
   test("tab bar shows action buttons per pane", async ({ page }) => {
     await gotoWorkspace(page, workspace.workspaceId);
     await assertSingleNewTabButton(page);
     await assertNewChatTileVisible(page);
     await assertNewTabMenuTriggerVisible(page);
   });
+
+  test("never shades the New tab button", async ({ page }) => {
+    await gotoWorkspace(page, workspace.workspaceId);
+
+    const tabRow = page.getByTestId("workspace-tabs-row").filter({ visible: true }).first();
+    const scrollArea = tabRow.getByTestId("workspace-tabs-scroll");
+    const newTabButtonInScroll = await scrollArea.getByTestId("workspace-new-tab-button").count();
+    const scrollShades = await tabRow
+      .locator('[data-testid^="workspace-tabs-scroll-shade-"]')
+      .count();
+
+    expect(newTabButtonInScroll === 0 || scrollShades === 0).toBe(true);
+    await expect(tabRow.getByTestId("workspace-new-tab-button")).toBeVisible();
+  });
 });
 
 // ═══════════════════════════════════════════════════════════════════════════
@@ -193,25 +378,35 @@ test.describe("Terminal title propagation", () => {
 // ═══════════════════════════════════════════════════════════════════════════
 
 test.describe("Tab transitions (no flash)", () => {
-  test("New agent tab transition has no blank intermediate tab state", async ({ page }) => {
-    await gotoWorkspace(page, workspace.workspaceId);
+  test("New agent tab transition has no blank intermediate tab state", async ({
+    page,
+    withWorkspace,
+  }) => {
+    const isolatedWorkspace = await withWorkspace({ prefix: "launcher-no-flash-" });
+    await isolatedWorkspace.navigateTo();
+    await pressNewTabShortcut(page);
+    await expect(
+      page.getByTestId("workspace-new-tab-panel").filter({ visible: true }),
+    ).toBeVisible();
 
-    // Sample tabs at high frequency across the transition
-    const snapshots = await sampleTabsDuringTransition(page, () => clickNewChat(page), 2_000, 30);
+    // Sample the single New → Agent replacement, not the separate action that
+    // creates the New tab in the first place.
+    const snapshots = await sampleTabsDuringTransition(page, async () => {
+      await page.getByTestId("workspace-new-tab-agent").filter({ visible: true }).first().click();
+    });
 
     // Every snapshot should have at least one tab — no blank/zero-tab frames
     for (const snapshot of snapshots) {
       expect(snapshot.length).toBeGreaterThanOrEqual(1);
     }
 
-    // Tab count should never spike excessively (no duplicate flash from add-then-remove).
-    // When running in-suite, previous tests may have created tabs on the shared workspace,
-    // so we allow +2 tolerance for accumulated state and React render batching.
-    const counts = snapshots.map((s) => s.length);
-    const maxCount = Math.max(...counts);
+    // Replacement is atomic: the set of identities changes once, while its size stays fixed.
+    const counts = snapshots.map((snapshot) => snapshot.length);
     const initialCount = counts[0] ?? 0;
 
-    expect(maxCount).toBeLessThanOrEqual(initialCount + 2);
+    expect(counts.every((count) => count === initialCount)).toBe(true);
+    expect(new Set(snapshots.map(tabIdentityKey)).size).toBeLessThanOrEqual(2);
+    await expectTabTitleFits(page, "New Agent", { min: 96, max: 160 });
   });
 
   test("Terminal transition completes within visual budget", async ({ page }) => {

packages/app/e2e/browser/mermaid-streaming.spec.ts

@@ -0,0 +1,70 @@
+import { test } from "../support/fixtures";
+import {
+  expectCompletedDiagram,
+  expectDiagramRemainsRenderedWhileStreaming,
+  expectDiagramWithLabels,
+  reloadConversation,
+  requestDiagram,
+  waitForDiagramTurnToComplete,
+} from "../support/helpers/diagram";
+import { openAgentRoute, seedMockAgentWorkspace } from "../support/helpers/mock-agent";
+
+const STREAMED_MERMAID = [
+  "```mermaid",
+  "flowchart LR",
+  "  Start --> Middle",
+  '  Middle --> Done["<i>Done</i>"]',
+  "  Middle --> Review",
+  "  Review --> Verify",
+  "  Verify --> Ship",
+  "  Ship --> Package",
+  "  Package --> Sign",
+  "  Sign --> Upload",
+  "  Upload --> Publish",
+  "  Publish --> Deploy",
+  "  Deploy --> Observe",
+  "  Observe --> Validate",
+  "  Validate --> Announce",
+  "  Announce --> Document",
+  "  Document --> Archive",
+  "  Archive --> Release",
+  "```",
+].join("\n");
+
+test("keeps a Mermaid diagram rendered while its message streams, completes, and reloads", async ({
+  page,
+}) => {
+  test.setTimeout(120_000);
+
+  const agent = await seedMockAgentWorkspace({
+    repoPrefix: "mermaid-streaming-",
+    title: "Streaming Mermaid",
+    featureValues: {
+      mockStreamingAssistantResponse: STREAMED_MERMAID,
+      mockStreamingAssistantIntervalMs: 150,
+    },
+  });
+  try {
+    await test.step("Open the conversation and request a diagram", async () => {
+      await openAgentRoute(page, agent);
+      await requestDiagram(agent);
+    });
+
+    await test.step("The diagram renders and stays rendered while tokens arrive", async () => {
+      await expectDiagramWithLabels(page, ["Start", "Middle"]);
+      await expectDiagramRemainsRenderedWhileStreaming(page);
+    });
+
+    await test.step("The completed diagram shows the final streamed content", async () => {
+      await waitForDiagramTurnToComplete(agent);
+      await expectCompletedDiagram(page, ["Start", "Done", "Release"]);
+    });
+
+    await test.step("The completed diagram remains rendered after reload", async () => {
+      await reloadConversation(page);
+      await expectCompletedDiagram(page, ["Start", "Done", "Release"]);
+    });
+  } finally {
+    await agent.cleanup();
+  }
+});

packages/app/e2e/browser/metadata-generation-settings.spec.ts

@@ -0,0 +1,110 @@
+import { expect, test, type Page } from "../support/fixtures";
+import { gotoAppShell, openSettings } from "../support/helpers/app";
+import { getServerId } from "../support/helpers/server-id";
+import { connectNewWorkspaceDaemonClient } from "../support/helpers/new-workspace";
+import {
+  expectSettingsHeader,
+  openHostSection,
+  openSettingsHost,
+} from "../support/helpers/settings";
+
+async function openMetadataGenerationSettings(page: Page) {
+  const serverId = getServerId();
+  await gotoAppShell(page);
+  await openSettings(page);
+  await openSettingsHost(page, serverId);
+  await openHostSection(page, serverId, "metadata");
+  await expectSettingsHeader(page, "Metadata");
+}
+
+async function openManualMetadataModelPicker(page: Page) {
+  await page.getByRole("button", { name: "Manual", exact: true }).click();
+  await page.getByRole("button", { name: /Select model/ }).click();
+}
+
+test("chooses a metadata model and can return to automatic selection", async ({
+  page,
+}, testInfo) => {
+  await openMetadataGenerationSettings(page);
+
+  await expect(page.getByText("Workspace titles", { exact: false })).toBeVisible();
+  await expect(page.getByRole("button", { name: "Automatic", exact: true })).toHaveAttribute(
+    "aria-selected",
+    "true",
+  );
+  await page.screenshot({
+    path: testInfo.outputPath("metadata-automatic.png"),
+    fullPage: true,
+  });
+
+  await openManualMetadataModelPicker(page);
+  await page.getByText("Mock Load Test", { exact: true }).click();
+  await expect(page.getByText("Ten second stream", { exact: true })).toBeVisible();
+  await page.screenshot({
+    path: testInfo.outputPath("metadata-model-picker.png"),
+    fullPage: true,
+  });
+  await page.getByText("Ten second stream", { exact: true }).click();
+
+  await expect(page.getByRole("button", { name: /Ten second stream/ })).toBeVisible();
+  await page.reload();
+  await expect(page.getByRole("button", { name: "Manual", exact: true })).toHaveAttribute(
+    "aria-selected",
+    "true",
+  );
+  await expect(page.getByRole("button", { name: /Ten second stream/ })).toBeVisible();
+  await page.screenshot({
+    path: testInfo.outputPath("metadata-manual-persisted.png"),
+    fullPage: true,
+  });
+
+  await page.getByRole("button", { name: "Automatic", exact: true }).click();
+  await page.reload();
+
+  await expect(page.getByRole("button", { name: "Automatic", exact: true })).toHaveAttribute(
+    "aria-selected",
+    "true",
+  );
+  await expect(page.getByRole("button", { name: /Ten second stream/ })).toHaveCount(0);
+});
+
+test("replaces only the first configured metadata model", async ({ page }) => {
+  const client = await connectNewWorkspaceDaemonClient({ ownProjects: false });
+  const previousConfig = await client.getDaemonConfig();
+
+  try {
+    await client.patchDaemonConfig({
+      metadataGeneration: {
+        providers: [
+          { provider: "mock", model: "five-minute-stream" },
+          { provider: "mock", model: "thirty-minute-stream" },
+        ],
+      },
+    });
+    await openMetadataGenerationSettings(page);
+
+    await expect(page.getByRole("button", { name: "Manual", exact: true })).toHaveAttribute(
+      "aria-selected",
+      "true",
+    );
+    await openManualMetadataModelPicker(page);
+    await page.getByText("Ten second stream", { exact: true }).click();
+
+    await expect
+      .poll(async () => (await client.getDaemonConfig()).config.metadataGeneration.providers)
+      .toEqual([
+        { provider: "mock", model: "ten-second-stream" },
+        { provider: "mock", model: "thirty-minute-stream" },
+      ]);
+  } finally {
+    try {
+      await client.patchDaemonConfig({
+        metadataGeneration: {
+          providers: previousConfig.config.metadataGeneration.providers,
+        },
+      });
+    } finally {
+      await client.close().catch(() => undefined);
+    }
+  }
+});

packages/app/e2e/browser/model-search-across-providers.spec.ts

@@ -0,0 +1,160 @@
+import { test } from "../support/fixtures";
+import {
+  closeModelPicker,
+  expectModelSearchEmptyState,
+  expectModelPickerHeight,
+  expectModelPickerWidth,
+  expectModelSearchResult,
+  expectPinnedProfilesHidden,
+  openModelPicker,
+  readModelPickerHeight,
+  readModelPickerWidth,
+  searchAllModels,
+  seedAgentProfiles,
+  seedModelProvider,
+  expectSearchResultsVirtualized,
+} from "../support/helpers/agent-profiles";
+import { expectComposerVisible } from "../support/helpers/composer";
+import { clickNewChat, gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+
+const MOCK_PROVIDER_LABEL = "Mock Load Test";
+
+// Its fast model deliberately carries the same label as the mock provider's, so
+// the only thing telling the two result rows apart is the provider name.
+const STUDIO = {
+  id: "mock-studio",
+  label: "Mock Studio",
+  models: [
+    { id: "studio-fast", label: "Ten second stream", description: "Studio quick pass" },
+    { id: "studio-deep", label: "Studio deep think", description: "Studio long pass" },
+  ],
+};
+
+// The cross-provider search lives on the picker's root view, and a profile is
+// what keeps the picker opening there.
+const ROOT_VIEW_PROFILE = {
+  id: "agent_profile_e2e_search_root",
+  name: "Search anchor",
+  provider: "mock",
+};
+
+const LARGE_CATALOG_SIZE = 200;
+const LARGE_CATALOG = {
+  id: "mock-bulk",
+  label: "Mock Bulk",
+  models: Array.from({ length: LARGE_CATALOG_SIZE }, (_, index) => ({
+    id: `bulk-${index}`,
+    label: `Bulk model ${index}`,
+    description: `Bulk search result ${index}`,
+  })),
+};
+
+test.describe("Cross-provider model search", () => {
+  test("one query over the picker root reaches every provider and names each result's provider", async ({
+    page,
+  }) => {
+    const provider = await seedModelProvider(STUDIO);
+    const profile = await seedAgentProfiles([ROOT_VIEW_PROFILE]);
+    const workspace = await seedWorkspace({ repoPrefix: "model-search-providers-" });
+
+    try {
+      await test.step("open a draft composer that can reach every provider", async () => {
+        await gotoWorkspace(page, workspace.workspaceId);
+        await clickNewChat(page);
+        await expectComposerVisible(page);
+        await openModelPicker(page);
+      });
+      const restingWidth = await readModelPickerWidth(page);
+
+      await test.step("one query returns the same model label from two providers", async () => {
+        await searchAllModels(page, "ten second stream");
+        await expectModelSearchResult(page, {
+          provider: "mock",
+          modelId: "ten-second-stream",
+          modelLabel: "Ten second stream",
+          providerLabel: MOCK_PROVIDER_LABEL,
+        });
+        await expectModelSearchResult(page, {
+          provider: STUDIO.id,
+          modelId: "studio-fast",
+          modelLabel: "Ten second stream",
+          providerLabel: STUDIO.label,
+        });
+        await expectModelPickerWidth(page, restingWidth);
+      });
+
+      await test.step("searching by provider name reaches that provider's own models", async () => {
+        await searchAllModels(page, "studio deep");
+        await expectModelSearchResult(page, {
+          provider: STUDIO.id,
+          modelId: "studio-deep",
+          modelLabel: "Studio deep think",
+          providerLabel: STUDIO.label,
+        });
+      });
+
+      await test.step("results replace the pinned profiles", async () => {
+        await expectPinnedProfilesHidden(page);
+      });
+
+      // Search falls back to subsequence matching, so "no matches" needs letters
+      // that cannot be picked out of a model label or description in order.
+      await test.step("a query with no matches repeats the query back", async () => {
+        await searchAllModels(page, "zzz");
+        await expectModelSearchEmptyState(page, "zzz");
+        await closeModelPicker(page);
+      });
+    } finally {
+      await workspace.cleanup();
+      await profile.restore();
+      await provider.restore();
+    }
+  });
+
+  test("desktop search keeps its frame stable and virtualizes a host-wide catalog", async ({
+    page,
+  }) => {
+    const provider = await seedModelProvider(LARGE_CATALOG);
+    const profile = await seedAgentProfiles([ROOT_VIEW_PROFILE]);
+    const workspace = await seedWorkspace({ repoPrefix: "model-search-large-catalog-" });
+
+    try {
+      await test.step("open the host-wide model picker", async () => {
+        await gotoWorkspace(page, workspace.workspaceId);
+        await clickNewChat(page);
+        await expectComposerVisible(page);
+        await openModelPicker(page);
+      });
+
+      const restingHeight = await readModelPickerHeight(page);
+      const restingWidth = await readModelPickerWidth(page);
+
+      await test.step("a broad query renders only the visible result window", async () => {
+        await searchAllModels(page, "bulk model");
+        await expectSearchResultsVirtualized(page, {
+          provider: LARGE_CATALOG.id,
+          total: LARGE_CATALOG_SIZE,
+        });
+        await expectModelPickerHeight(page, restingHeight);
+        await expectModelPickerWidth(page, restingWidth);
+      });
+
+      await test.step("narrowing to one result does not resize the picker", async () => {
+        await searchAllModels(page, "bulk model 199");
+        await expectModelSearchResult(page, {
+          provider: LARGE_CATALOG.id,
+          modelId: "bulk-199",
+          modelLabel: "Bulk model 199",
+          providerLabel: LARGE_CATALOG.label,
+        });
+        await expectModelPickerHeight(page, restingHeight);
+        await expectModelPickerWidth(page, restingWidth);
+      });
+    } finally {
+      await workspace.cleanup();
+      await profile.restore();
+      await provider.restore();
+    }
+  });
+});

packages/app/e2e/browser/new-tab-startup.spec.ts

@@ -0,0 +1,49 @@
+import { expect } from "@playwright/test";
+import { test } from "../support/fixtures";
+import { gotoWorkspace } from "../support/helpers/launcher";
+import { seedWorkspace } from "../support/helpers/seed-client";
+
+test("a workspace with an existing agent never presents New during startup", async ({ page }) => {
+  const workspace = await seedWorkspace({ repoPrefix: "new-tab-startup-agent-" });
+  const agent = await workspace.client.createAgent({
+    provider: "mock",
+    cwd: workspace.repoPath,
+    workspaceId: workspace.workspaceId,
+    title: "Existing startup agent",
+    modeId: "load-test",
+    model: "ten-second-stream",
+    initialPrompt: "ready",
+  });
+
+  try {
+    await page.addInitScript(() => {
+      Reflect.set(globalThis, "__paseoSawVisibleNewTab", false);
+      function recordVisibleNewTab() {
+        const panel = document.querySelector('[data-testid="workspace-new-tab-panel"]');
+        if (panel?.getClientRects().length) {
+          Reflect.set(globalThis, "__paseoSawVisibleNewTab", true);
+        }
+      }
+      window.addEventListener("DOMContentLoaded", () => {
+        new MutationObserver(recordVisibleNewTab).observe(document.documentElement, {
+          childList: true,
+          subtree: true,
+        });
+        recordVisibleNewTab();
+      });
+    });
+
+    await gotoWorkspace(page, workspace.workspaceId);
+    await expect(
+      page.getByTestId(`workspace-tab-agent_${agent.id}`).filter({ visible: true }),
+    ).toHaveAttribute("aria-selected", "true", { timeout: 30_000 });
+    await expect(page.getByTestId("workspace-new-tab-panel").filter({ visible: true })).toHaveCount(
+      0,
+    );
+    expect(await page.evaluate(() => Reflect.get(globalThis, "__paseoSawVisibleNewTab"))).toBe(
+      false,
+    );
+  } finally {
+    await workspace.cleanup();
+  }
+});

packages/app/e2e/browser/new-workspace-codex-mode-preferences.spec.ts

@@ -1,4 +1,5 @@
 import { expect, test, type Page } from "../support/fixtures";
+import type { FormPreferences } from "@/create-agent-preferences/preferences";
 import { gotoAppShell } from "../support/helpers/app";
 import { daemonWsRoutePattern } from "../support/helpers/daemon-port";
 import { openAgentRoute } from "../support/helpers/mock-agent";
@@ -10,7 +11,6 @@ import {
 import { expectNoTruncation } from "../support/helpers/no-truncation";
 import { escapeRegex } from "../support/helpers/regex";
 import { seedWorkspace } from "../support/helpers/seed-client";
-import { getServerId } from "../support/helpers/server-id";
 import { waitForSidebarHydration } from "../support/helpers/workspace-ui";
 
 const CREATE_AGENT_PREFERENCES_KEY = "@paseo:create-agent-preferences";
@@ -49,13 +49,12 @@ function getSessionMessage(message: WebSocketMessage): Record<string, unknown> |
   return maybeEnvelope.message as Record<string, unknown>;
 }
 
-async function seedCodexDefaultPermissionPreferences(page: Page, serverId: string): Promise<void> {
+async function seedCodexDefaultPermissionPreferences(page: Page): Promise<void> {
   await page.addInitScript(
-    ({ preferencesKey, serverId: seededServerId }) => {
+    ({ preferencesKey }) => {
       localStorage.setItem(
         preferencesKey,
         JSON.stringify({
-          serverId: seededServerId,
           provider: "codex",
           providerPreferences: {
             codex: {
@@ -69,10 +68,10 @@ async function seedCodexDefaultPermissionPreferences(page: Page, serverId: strin
               model: "ten-second-stream",
             },
           },
-        }),
+        } satisfies FormPreferences),
       );
     },
-    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY, serverId },
+    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY },
   );
 }
 
@@ -155,10 +154,9 @@ test.describe("New workspace Codex mode preferences", () => {
   test("keeps Full Access as the global Codex mode after the workspace draft auto-submit handoff", async ({
     page,
   }) => {
-    const serverId = getServerId();
     const seeded = await seedWorkspace({ repoPrefix: "codex-mode-preferences-" });
     const createAgentRecorder = await recordAndBlockCreateAgentRequests(page);
-    await seedCodexDefaultPermissionPreferences(page, serverId);
+    await seedCodexDefaultPermissionPreferences(page);
 
     try {
       await gotoAppShell(page);
@@ -194,9 +192,8 @@ test.describe("New workspace Codex mode preferences", () => {
   });
 
   test("uses the live Codex agent mode as the next New Workspace default", async ({ page }) => {
-    const serverId = getServerId();
     const seeded = await seedWorkspace({ repoPrefix: "codex-live-mode-preferences-" });
-    await seedCodexDefaultPermissionPreferences(page, serverId);
+    await seedCodexDefaultPermissionPreferences(page);
 
     try {
       const agent = await seeded.client.createAgent({

packages/app/e2e/browser/new-workspace-composer-draft.spec.ts

@@ -1,4 +1,4 @@
-import { test } from "../support/fixtures";
+import { expect, test } from "../support/fixtures";
 import { gotoAppShell } from "../support/helpers/app";
 import { getE2EDaemonPort } from "../support/helpers/daemon-port";
 import {
@@ -85,4 +85,52 @@ test.describe("New workspace composer draft", () => {
       await project.cleanup();
     }
   });
+
+  test("does not restore a submitted draft after deferred publication", async ({ page }) => {
+    const project = await seedWorkspace({ repoPrefix: "new-workspace-submitted-draft-" });
+
+    try {
+      await gotoAppShell(page);
+      await waitForSidebarHydration(page);
+      await openNewWorkspaceComposer(page, {
+        projectKey: project.projectKey,
+        projectDisplayName: project.projectDisplayName,
+      });
+
+      const composer = page.getByRole("textbox", { name: "Message agent..." });
+      const createButton = page.getByTestId("workspace-create-submit");
+      await expect(composer).toBeEditable({ timeout: 30_000 });
+      await expect(createButton).toBeEnabled({ timeout: 30_000 });
+
+      await composer.evaluate((element, draft) => {
+        if (!(element instanceof HTMLTextAreaElement)) {
+          throw new Error("Composer input is not a textarea");
+        }
+        const valueSetter = Object.getOwnPropertyDescriptor(
+          HTMLTextAreaElement.prototype,
+          "value",
+        )?.set;
+        if (!valueSetter) throw new Error("Textarea value setter is unavailable");
+        valueSetter.call(element, draft);
+        element.dispatchEvent(
+          new InputEvent("input", {
+            bubbles: true,
+            data: draft,
+            inputType: "insertText",
+          }),
+        );
+        const button = document.querySelector('[data-testid="workspace-create-submit"]');
+        if (!(button instanceof HTMLElement)) {
+          throw new Error("Create button is unavailable");
+        }
+        button.click();
+      }, DRAFT);
+
+      await page.waitForURL((url) => url.pathname.includes("/workspace/"), { timeout: 30_000 });
+      await openGlobalNewWorkspaceComposer(page);
+      await expectNewWorkspaceDraft(page, "");
+    } finally {
+      await project.cleanup();
+    }
+  });
 });

packages/app/e2e/browser/new-workspace-launch-memory.spec.ts

@@ -64,6 +64,7 @@ test.describe("New workspace: launch target memory", () => {
     await test.step("launching a terminal profile preselects it the next time New workspace opens", async () => {
       await openNewWorkspace();
       await selectLaunchOption(page, PROMPT_PROFILE.id);
+      await expectTerminalLaunchSelected(page, PROMPT_PROFILE.name);
       await fillTerminalPrompt(page, "remember me");
       await submitTerminalLaunch(page);
       await expectWorkspaceOpensWithTerminalTab(page);
@@ -88,6 +89,7 @@ test.describe("New workspace: launch target memory", () => {
 
     await test.step("re-selecting the terminal profile, then removing it daemon-side, falls back to Chat rather than a dead entry", async () => {
       await selectLaunchOption(page, PROMPT_PROFILE.id);
+      await expectTerminalLaunchSelected(page, PROMPT_PROFILE.name);
       await fillTerminalPrompt(page, "second run");
       await submitTerminalLaunch(page);
       await expectWorkspaceOpensWithTerminalTab(page);

packages/app/e2e/browser/new-workspace-mode-cycle-safety.spec.ts

@@ -1,4 +1,5 @@
 import { expect, test, type Page } from "../support/fixtures";
+import type { FormPreferences } from "@/create-agent-preferences/preferences";
 import { daemonWsRoutePattern } from "../support/helpers/daemon-port";
 import { openAgentRoute } from "../support/helpers/mock-agent";
 import {
@@ -6,7 +7,6 @@ import {
   selectNewWorkspaceProject,
 } from "../support/helpers/new-workspace";
 import { seedWorkspace } from "../support/helpers/seed-client";
-import { getServerId } from "../support/helpers/server-id";
 
 const CREATE_AGENT_PREFERENCES_KEY = "@paseo:create-agent-preferences";
 
@@ -62,13 +62,12 @@ async function recordSetAgentModeRequests(page: Page): Promise<{
   };
 }
 
-async function seedCodexDefaultPreferences(page: Page, serverId: string): Promise<void> {
+async function seedCodexDefaultPreferences(page: Page): Promise<void> {
   await page.addInitScript(
-    ({ preferencesKey, serverId: seededServerId }) => {
+    ({ preferencesKey }) => {
       localStorage.setItem(
         preferencesKey,
         JSON.stringify({
-          serverId: seededServerId,
           provider: "codex",
           providerPreferences: {
             codex: {
@@ -78,10 +77,10 @@ async function seedCodexDefaultPreferences(page: Page, serverId: string): Promis
             },
             mock: { model: "ten-second-stream" },
           },
-        }),
+        } satisfies FormPreferences),
       );
     },
-    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY, serverId },
+    { preferencesKey: CREATE_AGENT_PREFERENCES_KEY },
   );
 }
 
@@ -104,9 +103,8 @@ test.describe("New Workspace mode cycle safety", () => {
   // control and silently change that (possibly running) agent's mode — e.g. into a
   // permissive/bypass mode. See use-keyboard-action-handler.ts.
   test("Shift+Tab in New Workspace never changes a backgrounded agent's mode", async ({ page }) => {
-    const serverId = getServerId();
     const seeded = await seedWorkspace({ repoPrefix: "mode-cycle-safety-" });
-    await seedCodexDefaultPreferences(page, serverId);
+    await seedCodexDefaultPreferences(page);
     const modeRequests = await recordSetAgentModeRequests(page);
 
     try {

packages/app/e2e/browser/permission-steer.real.spec.ts

@@ -0,0 +1,83 @@
+import { mkdtempSync, realpathSync } from "node:fs";
+import { tmpdir } from "node:os";
+import path from "node:path";
+import { expect, test } from "../support/fixtures";
+import { submitMessage } from "../support/helpers/composer";
+import { waitForPermissionPrompt } from "../support/helpers/permissions";
+import { cleanupRewindFlow, type AgentHandle, launchAgent } from "../support/helpers/rewind-flow";
+
+const scenarios = [
+  {
+    provider: "claude" as const,
+    providerConfig: { model: "haiku", modeId: "plan" },
+    planPrompt:
+      "Do not ask questions. Produce a concise plan to append the line 'Hello' to README.md, then present it for approval. Do not implement it.",
+    planText: "README.md",
+    steerPrompt: "Do not implement anything. Reply exactly CLAUDE_STEER_RECEIVED and stop.",
+    reply: "CLAUDE_STEER_RECEIVED",
+  },
+  {
+    provider: "codex" as const,
+    providerConfig: { featureValues: { plan_mode: true } },
+    planPrompt:
+      "Produce a concise implementation plan with exactly these steps: Inspect permission steering; Implement permission steering; Verify permission steering. Do not implement it.",
+    planText: "Inspect permission steering",
+    steerPrompt: "Do not produce another plan. Reply exactly CODEX_STEER_RECEIVED and stop.",
+    reply: "CODEX_STEER_RECEIVED",
+  },
+];
+
+test.describe("composer steer supersedes plan approval", () => {
+  for (const scenario of scenarios) {
+    test(`${scenario.provider} keeps the rejected plan in the timeline`, async ({
+      page,
+    }, testInfo) => {
+      test.setTimeout(420_000);
+      const cwd = realpathSync(
+        mkdtempSync(path.join(tmpdir(), `paseo-permission-steer-${scenario.provider}-`)),
+      );
+      let handle: AgentHandle | undefined;
+
+      try {
+        handle = await launchAgent({
+          page,
+          provider: scenario.provider,
+          cwd,
+          mode: "full-access",
+          providerConfig: scenario.providerConfig,
+        });
+        await submitMessage(page, scenario.planPrompt);
+        await waitForPermissionPrompt(page, 180_000);
+
+        const pendingPlan = page.getByTestId("permission-plan-card");
+        await expect(pendingPlan).toContainText(scenario.planText);
+        const pendingScreenshot = testInfo.outputPath(`${scenario.provider}-pending-plan.png`);
+        await page.screenshot({ path: pendingScreenshot });
+        await testInfo.attach(`${scenario.provider} pending plan`, {
+          path: pendingScreenshot,
+          contentType: "image/png",
+        });
+
+        await submitMessage(page, scenario.steerPrompt);
+        await expect(pendingPlan).toHaveCount(0, { timeout: 30_000 });
+        const rejectedPlan = page.getByTestId("timeline-plan-card");
+        await expect(rejectedPlan).toBeVisible({ timeout: 30_000 });
+        await expect(rejectedPlan).toContainText(scenario.planText);
+        await expect(
+          page.getByTestId("assistant-message").filter({ hasText: scenario.reply }),
+        ).toBeVisible({
+          timeout: 180_000,
+        });
+
+        const rejectedScreenshot = testInfo.outputPath(`${scenario.provider}-rejected-plan.png`);
+        await page.screenshot({ path: rejectedScreenshot });
+        await testInfo.attach(`${scenario.provider} rejected plan`, {
+          path: rejectedScreenshot,
+          contentType: "image/png",
+        });
+      } finally {
+        await cleanupRewindFlow({ handle, cwd });
+      }
+    });
+  }
+});

packages/app/e2e/browser/plugin-management.spec.ts

@@ -0,0 +1,209 @@
+import { mkdtemp, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import path from "node:path";
+import { expect, test, type Page } from "../support/fixtures";
+import { gotoAppShell, openSettings } from "../support/helpers/app";
+import { getServerId } from "../support/helpers/server-id";
+import { connectNewWorkspaceDaemonClient } from "../support/helpers/new-workspace";
+import {
+  expectSettingsHeader,
+  openHostSection,
+  openSettingsHost,
+} from "../support/helpers/settings";
+
+function observePluginCatalog(page: Page) {
+  let responses = 0;
+  page.on("websocket", (socket) => {
+    socket.on("framereceived", ({ payload }) => {
+      if (typeof payload !== "string") return;
+      try {
+        const envelope = JSON.parse(payload) as {
+          type?: unknown;
+          message?: { type?: unknown };
+        };
+        const message = envelope.type === "session" ? envelope.message : envelope;
+        if (message?.type === "plugin.catalog.get.response") responses += 1;
+      } catch {
+        return;
+      }
+    });
+  });
+  return {
+    waitForInitialFetch: async () => {
+      await expect.poll(() => responses).toBeGreaterThan(0);
+    },
+  };
+}
+
+function pluginSource(title: string, renderError?: string): string {
+  return `import React from "react";
+import { Text } from "react-native";
+export default function contribute(plugin) {
+  function Surface() {
+    ${renderError ? `throw new Error(${JSON.stringify(renderError)});` : ""}
+    const cleanups = Number(globalThis.sessionStorage?.getItem("paseo-plugin-cleanups") || "0");
+    return <Text>${title} cleanup {cleanups}</Text>;
+  }
+  plugin.addSurface("main", Surface);
+  plugin.addSidebarItem({ id: "main", title: ${JSON.stringify(title)}, icon: "Blocks", surface: "main" });
+  ${
+    renderError
+      ? `function HealthySurface() { return <Text>Healthy contribution</Text>; }
+  plugin.addSurface("healthy", HealthySurface);
+  plugin.addSidebarItem({ id: "healthy", title: "Healthy surface", icon: "Blocks", surface: "healthy" });`
+      : ""
+  }
+  return () => {
+    const storage = globalThis.sessionStorage;
+    if (storage) {
+      const cleanups = Number(storage.getItem("paseo-plugin-cleanups") || "0");
+      storage.setItem("paseo-plugin-cleanups", String(cleanups + 1));
+    }
+  };
+}`;
+}
+
+async function openPluginSettings(page: Page): Promise<void> {
+  const serverId = getServerId();
+  await openSettings(page);
+  await openSettingsHost(page, serverId);
+  await openHostSection(page, serverId, "plugins");
+  await expectSettingsHeader(page, "Plugins");
+}
+
+async function reloadPlugin(page: Page): Promise<void> {
+  await page.getByRole("button", { name: "Reload", exact: true }).click();
+  await expect(page.getByText("Reloaded e2e-plugin", { exact: true })).toBeVisible();
+}
+
+async function expectContributionBody(page: Page, expectedBody: string): Promise<void> {
+  const body = page.getByText(expectedBody, { exact: true }).filter({ visible: true });
+  await expect(body).toHaveCount(1);
+  await expect(body).toBeVisible();
+}
+
+async function openContribution(page: Page, title: string, expectedBody: string): Promise<void> {
+  await page.getByRole("button", { name: title, exact: true }).click();
+  await expectContributionBody(page, expectedBody);
+}
+
+async function leavePluginSettings(page: Page): Promise<void> {
+  await page.getByRole("button", { name: "Back", exact: true }).click();
+  await expect(page).not.toHaveURL(/\/settings\//);
+}
+
+async function openContributionFromSettings(
+  page: Page,
+  title: string,
+  expectedBody: string,
+): Promise<void> {
+  await leavePluginSettings(page);
+  await openContribution(page, title, expectedBody);
+}
+
+async function reloadActiveContribution(
+  page: Page,
+  client: Awaited<ReturnType<typeof connectNewWorkspaceDaemonClient>>,
+  expectedBody: string,
+): Promise<void> {
+  await client.reloadPlugin("e2e-plugin");
+  await expectContributionBody(page, expectedBody);
+}
+
+async function expectContributionRemoved(page: Page, title: string): Promise<void> {
+  await expect(page.getByRole("button", { name: title, exact: true })).not.toBeVisible();
+}
+
+test("installs, reloads, recovers, disables, and removes a trusted local plugin", async ({
+  page,
+}) => {
+  const directory = await mkdtemp(path.join(tmpdir(), "paseo-plugin-e2e-"));
+  const client = await connectNewWorkspaceDaemonClient({ ownProjects: false });
+  const previous = await client.getDaemonConfig();
+  await writeFile(path.join(directory, "paseo-plugin.json"), JSON.stringify({ id: "e2e-plugin" }));
+  await writeFile(path.join(directory, "index.tsx"), pluginSource("Plugin v1"));
+
+  try {
+    const catalog = observePluginCatalog(page);
+    await gotoAppShell(page);
+    await openPluginSettings(page);
+    await catalog.waitForInitialFetch();
+    await client.patchDaemonConfig({ pluginsEnabled: false });
+    await page.getByRole("switch", { name: "Enable plugins" }).click();
+    await expect(page.getByText("Plugins enabled", { exact: true })).toBeVisible();
+    await page.getByLabel("Plugin directory").fill(directory);
+    await page.getByLabel("Plugin installation ID").focus();
+    await expect(page.getByLabel("Plugin installation ID")).toHaveValue("e2e-plugin");
+    await page.getByRole("button", { name: "Install directory" }).click();
+    await expect(page.getByText("Installed e2e-plugin", { exact: true })).toBeVisible();
+    await expect(page.getByLabel("e2e-plugin running")).toBeVisible();
+    await openContributionFromSettings(page, "Plugin v1", "Plugin v1 cleanup 0");
+
+    await openPluginSettings(page);
+    await reloadPlugin(page);
+    await openContributionFromSettings(page, "Plugin v1", "Plugin v1 cleanup 1");
+
+    await writeFile(path.join(directory, "index.tsx"), pluginSource("Plugin v2"));
+    await openPluginSettings(page);
+    await reloadPlugin(page);
+    await openContributionFromSettings(page, "Plugin v2", "Plugin v2 cleanup 2");
+    await expect(page.getByRole("button", { name: "Plugin v1", exact: true })).not.toBeVisible();
+
+    await writeFile(
+      path.join(directory, "index.tsx"),
+      pluginSource("Broken surface", "render exploded"),
+    );
+    await openPluginSettings(page);
+    await reloadPlugin(page);
+    await openContributionFromSettings(page, "Broken surface", "Plugin failed: render exploded");
+    await openContribution(page, "Healthy surface", "Healthy contribution");
+    await openContribution(page, "Broken surface", "Plugin failed: render exploded");
+
+    await writeFile(path.join(directory, "index.tsx"), pluginSource("Recovered surface"));
+    await reloadActiveContribution(page, client, "Recovered surface cleanup 4");
+    await expect(
+      page.getByRole("button", { name: "Broken surface", exact: true }),
+    ).not.toBeVisible();
+
+    await writeFile(path.join(directory, "index.tsx"), "export default broken syntax !!!");
+    await openPluginSettings(page);
+    await page.getByRole("button", { name: "Reload", exact: true }).click();
+    await expect(page.getByTestId("plugin-management-feedback")).toContainText("Request failed");
+    await expect(page.getByLabel("e2e-plugin failed")).toBeVisible();
+    await leavePluginSettings(page);
+    await expectContributionRemoved(page, "Recovered surface");
+
+    await writeFile(path.join(directory, "index.tsx"), pluginSource("Plugin v3"));
+    await openPluginSettings(page);
+    await page.getByRole("button", { name: "Reload", exact: true }).click();
+    await expect(page.getByLabel("e2e-plugin running")).toBeVisible();
+    await openContributionFromSettings(page, "Plugin v3", "Plugin v3 cleanup 5");
+
+    await openPluginSettings(page);
+    await page.getByRole("button", { name: "Disable", exact: true }).click();
+    await expect(page.getByLabel("e2e-plugin disabled")).toBeVisible();
+    await leavePluginSettings(page);
+    await expectContributionRemoved(page, "Plugin v3");
+
+    await openPluginSettings(page);
+    await page.getByRole("button", { name: "Enable", exact: true }).click();
+    await expect(page.getByLabel("e2e-plugin running")).toBeVisible();
+    await openContributionFromSettings(page, "Plugin v3", "Plugin v3 cleanup 6");
+
+    await openPluginSettings(page);
+    page.once("dialog", (dialog) => dialog.accept());
+    await page.getByRole("button", { name: "Remove", exact: true }).click();
+    await expect(page.getByText("No plugins configured", { exact: true })).toBeVisible();
+
+    await page.getByLabel("Plugin directory").fill(directory);
+    await page.getByRole("button", { name: "Install directory" }).click();
+    await expect(page.getByLabel("e2e-plugin running")).toBeVisible();
+    await gotoAppShell(page);
+    await openContribution(page, "Plugin v3", "Plugin v3 cleanup 7");
+  } finally {
+    await client.removePlugin("e2e-plugin").catch(() => undefined);
+    await client.patchDaemonConfig({ pluginsEnabled: previous.config.pluginsEnabled ?? false });
+    await client.close().catch(() => undefined);
+    await rm(directory, { recursive: true, force: true });
+  }
+});

packages/app/e2e/browser/plugin-theme.spec.ts

@@ -0,0 +1,118 @@
+import { mkdtemp, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import path from "node:path";
+import { expect, test } from "../support/fixtures";
+import { connectNewWorkspaceDaemonClient } from "../support/helpers/new-workspace";
+import { openSettingsSection } from "../support/helpers/settings";
+
+const PLUGIN_ID = "plugin-theme-e2e";
+
+// Catppuccin Mocha: base, text, surface0, surface1, mauve, subtext0, overlay0.
+const PLUGIN_SOURCE = `export default function contribute(plugin) {
+  plugin.addTheme({
+    id: "mocha",
+    name: "Catppuccin Mocha",
+    appearance: "dark",
+    colors: {
+      background: "#1e1e2e",
+      foreground: "#cdd6f4",
+      raised: "#313244",
+      control: "#45475a",
+      border: "#45475a",
+      accent: "#cba6f7",
+      mutedForeground: "#a6adc8",
+      ring: "#6c7086",
+    },
+  });
+  plugin.addTheme({
+    id: "latte",
+    name: "Catppuccin Latte",
+    appearance: "light",
+    colors: {
+      background: "#eff1f5",
+      foreground: "#4c4f69",
+      raised: "#e6e9ef",
+      control: "#dce0e8",
+      border: "#ccd0da",
+      accent: "#8839ef",
+      mutedForeground: "#6c6f85",
+      ring: "#9ca0b0",
+    },
+  });
+  return () => {};
+}`;
+
+// settingsStyles.sectionHeaderTitle paints from foregroundMuted, so the section heading proves the
+// contributed palette reached the semantic tokens rather than just the swatch.
+const MOCHA_MUTED_FOREGROUND = "rgb(166, 173, 200)";
+const LATTE_MUTED_FOREGROUND = "rgb(108, 111, 133)";
+
+test("applies a contributed theme and falls back when its plugin is gone", async ({
+  page,
+}, testInfo) => {
+  const directory = await mkdtemp(path.join(tmpdir(), "paseo-plugin-theme-e2e-"));
+  const client = await connectNewWorkspaceDaemonClient({ ownProjects: false });
+  const previousConfig = await client.getDaemonConfig();
+  await writeFile(path.join(directory, "paseo-plugin.json"), JSON.stringify({ id: PLUGIN_ID }));
+  await writeFile(path.join(directory, "index.ts"), PLUGIN_SOURCE);
+
+  try {
+    await client.patchDaemonConfig({ pluginsEnabled: true });
+    await client.installDirectoryPlugin(directory);
+    await page.goto("/settings");
+    await expect(page.getByTestId("settings-sidebar")).toBeVisible();
+    await openSettingsSection(page, "appearance");
+
+    const sectionTitle = page.getByText("Theme", { exact: true }).first();
+    await page.getByLabel("Theme: System", { exact: true }).click();
+    const mochaItem = page.getByText("Catppuccin Mocha", { exact: true });
+    await expect(mochaItem).toBeVisible({ timeout: 30_000 });
+    await page.screenshot({
+      path: testInfo.outputPath("plugin-theme-picker.png"),
+      animations: "disabled",
+      fullPage: true,
+    });
+
+    await test.step("a contributed light theme uses the light palette", async () => {
+      await page.getByText("Catppuccin Latte", { exact: true }).click();
+      await expect(page.getByLabel("Theme: Catppuccin Latte", { exact: true })).toBeVisible();
+      await expect(sectionTitle).toHaveCSS("color", LATTE_MUTED_FOREGROUND);
+      await page.getByLabel("Theme: Catppuccin Latte", { exact: true }).click();
+    });
+
+    await mochaItem.click();
+    await expect(page.getByLabel("Theme: Catppuccin Mocha", { exact: true })).toBeVisible();
+    await expect(sectionTitle).toHaveCSS("color", MOCHA_MUTED_FOREGROUND);
+    await page.screenshot({
+      path: testInfo.outputPath("plugin-theme-applied.png"),
+      fullPage: true,
+    });
+
+    await test.step("the selection survives a reload", async () => {
+      await page.reload();
+      await expect(page.getByLabel("Theme: Catppuccin Mocha", { exact: true })).toBeVisible({
+        timeout: 30_000,
+      });
+      await expect(sectionTitle).toHaveCSS("color", MOCHA_MUTED_FOREGROUND);
+    });
+
+    await test.step("removing the plugin falls back to the default theme", async () => {
+      await client.removePlugin(PLUGIN_ID);
+      await expect(page.getByLabel("Theme: System", { exact: true })).toBeVisible({
+        timeout: 30_000,
+      });
+      await expect(sectionTitle).not.toHaveCSS("color", MOCHA_MUTED_FOREGROUND);
+      await page.screenshot({
+        path: testInfo.outputPath("plugin-theme-fallback.png"),
+        fullPage: true,
+      });
+    });
+  } finally {
+    await client.removePlugin(PLUGIN_ID).catch(() => undefined);
+    await client
+      .patchDaemonConfig({ pluginsEnabled: previousConfig.config.pluginsEnabled ?? false })
+      .catch(() => undefined);
+    await client.close().catch(() => undefined);
+    await rm(directory, { recursive: true, force: true });
+  }
+});

packages/app/e2e/browser/plugin-workspace-panels.spec.ts

@@ -0,0 +1,247 @@
+import { mkdtemp, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import path from "node:path";
+import type { TestInfo } from "@playwright/test";
+import { expect, test, type Page } from "../support/fixtures";
+import { gotoAppShell } from "../support/helpers/app";
+import { openCommandCenter } from "../support/helpers/command-center";
+import { addConnectedHostAndReload } from "../support/helpers/hosts";
+import { startIsolatedHostDaemon } from "../support/helpers/isolated-host-daemon";
+import { buildAgentRoute } from "../support/helpers/mock-agent";
+import { connectNewWorkspaceDaemonClient } from "../support/helpers/new-workspace";
+import { seedWorkspace } from "../support/helpers/seed-client";
+import { getServerId } from "../support/helpers/server-id";
+import {
+  closeMobileAgentSidebar,
+  expectMobileAgentSidebarHidden,
+} from "../support/helpers/sidebar";
+import {
+  switchWorkspaceViaSidebar,
+  waitForWorkspaceInSidebar,
+} from "../support/helpers/workspace-ui";
+
+const PLUGIN_ID = "workspace-panel-e2e";
+const WIDE_VIEWPORT = { width: 1280, height: 900 };
+const COMPACT_VIEWPORT = { width: 390, height: 844 };
+
+function isSettledWorkspaceUrl(url: URL): boolean {
+  return url.pathname.includes("/workspace/") && !url.searchParams.has("open");
+}
+
+function pluginSource(): string {
+  return `import React, { useRef } from "react";
+import { Text, View } from "react-native";
+import { useAgent, useWorkspace } from "@getpaseo/plugin";
+
+function WorkspacePanel({ workspaceId, host, layout }) {
+  const workspace = useWorkspace(workspaceId, (value) => ({ id: value.id }));
+  const renderCount = useRef(0);
+  renderCount.current += 1;
+  return <View><Text>Workspace bridge {workspace?.id}</Text><Text>Workspace renders {renderCount.current}</Text><Text>Host {host.id}</Text><Text>Layout {layout.compact ? "compact" : "wide"}</Text></View>;
+}
+
+function AgentPanel({ workspaceId, agentId, host, layout }) {
+  const workspace = useWorkspace(workspaceId, (value) => ({ id: value.id }));
+  const agent = useAgent(agentId, (value) => ({ id: value.id }));
+  return <View><Text>Agent bridge {agent?.id}</Text><Text>Workspace {workspace?.id}</Text><Text>Host {host.id}</Text><Text>Layout {layout.compact ? "compact" : "wide"}</Text></View>;
+}
+
+function DirectCollisionSurface() {
+  return <View><Text>Direct collision surface</Text></View>;
+}
+
+function SidebarCollisionSurface() {
+  return <View><Text>Sidebar collision surface</Text></View>;
+}
+
+export default function contribute(plugin) {
+  plugin.addSurface("collision", DirectCollisionSurface);
+  plugin.addSurface("sidebar-destination", SidebarCollisionSurface);
+  plugin.addSidebarItem({ id: "collision", title: "Collision sidebar", icon: "Blocks", surface: "sidebar-destination" });
+  plugin.addWorkspacePanel({ id: "workspace", title: "Workspace inspector", icon: "PanelsTopLeft", context: "workspace", Component: WorkspacePanel });
+  plugin.addWorkspacePanel({ id: "agent", title: "Agent inspector", icon: "PanelTop", context: "agent", Component: AgentPanel });
+  plugin.addCommandCenterItem({ id: "global", title: "Plugin global action", icon: "Blocks", context: "global", onSelect() {} });
+  plugin.addCommandCenterItem({ id: "surface", title: "Open direct collision surface", icon: "Blocks", context: "workspace", onSelect({ openSurface }) { openSurface("collision"); } });
+  plugin.addCommandCenterItem({ id: "workspace", title: "Open plugin workspace", icon: "PanelsTopLeft", context: "workspace", onSelect({ openPanel }) { openPanel("workspace"); } });
+  plugin.addCommandCenterItem({ id: "agent", title: "Open plugin agent", icon: "PanelTop", context: "agent", onSelect({ openPanel }) { openPanel("agent"); } });
+  return () => {};
+}`;
+}
+
+async function searchCommands(page: Page, query: string) {
+  const panel = await openCommandCenter(page);
+  await panel.getByTestId("command-center-input").fill(query);
+  return panel;
+}
+
+async function expectCommandAvailable(page: Page, title: string): Promise<void> {
+  const panel = await searchCommands(page, title);
+  await expect(panel.getByRole("button", { name: title, exact: true })).toBeVisible();
+  await page.keyboard.press("Escape");
+}
+
+async function expectCommandUnavailable(page: Page, title: string): Promise<void> {
+  const panel = await searchCommands(page, title);
+  await expect(panel.getByRole("button", { name: title, exact: true })).toHaveCount(0);
+  await page.keyboard.press("Escape");
+}
+
+async function runCommand(page: Page, title: string): Promise<void> {
+  const panel = await searchCommands(page, title);
+  await panel.getByRole("button", { name: title, exact: true }).click();
+  await expect(panel).not.toBeVisible();
+}
+
+async function openCompactSidebar(page: Page): Promise<void> {
+  await page.getByRole("button", { name: "Open menu", exact: true }).click();
+  await expect(page.getByTestId("sidebar-command-center-search")).toBeVisible();
+}
+
+async function capture(page: Page, testInfo: TestInfo, name: string): Promise<void> {
+  const screenshot = testInfo.outputPath(`${name}.png`);
+  await page.screenshot({ path: screenshot });
+  await testInfo.attach(name, { path: screenshot, contentType: "image/png" });
+}
+
+test.describe("plugin workspace panels and Command Center", () => {
+  test.describe.configure({ timeout: 240_000 });
+
+  test("follows workspace, agent, host, compact, and unavailable state", async ({
+    page,
+  }, testInfo) => {
+    const directory = await mkdtemp(path.join(tmpdir(), "paseo-plugin-workspace-panel-e2e-"));
+    const primaryClient = await connectNewWorkspaceDaemonClient({ ownProjects: false });
+    const previousConfig = await primaryClient.getDaemonConfig();
+    const primary = await seedWorkspace({ repoPrefix: "plugin-panel-primary-" });
+    const secondaryDaemon = await startIsolatedHostDaemon("plugin-panel-secondary");
+    const secondary = await seedWorkspace({
+      repoPrefix: "plugin-panel-secondary-",
+      port: secondaryDaemon.port,
+    });
+    await writeFile(path.join(directory, "paseo-plugin.json"), JSON.stringify({ id: PLUGIN_ID }));
+    await writeFile(path.join(directory, "index.tsx"), pluginSource());
+
+    try {
+      await primaryClient.patchDaemonConfig({ pluginsEnabled: true });
+      await primaryClient.installDirectoryPlugin(directory);
+      await page.setViewportSize(WIDE_VIEWPORT);
+      await gotoAppShell(page);
+      await addConnectedHostAndReload(page, {
+        serverId: secondaryDaemon.serverId,
+        label: "Secondary plugin host",
+        port: secondaryDaemon.port,
+        primaryLabel: "Primary plugin host",
+      });
+      await waitForWorkspaceInSidebar(page, {
+        serverId: getServerId(),
+        workspaceId: primary.workspaceId,
+      });
+      await waitForWorkspaceInSidebar(page, {
+        serverId: secondaryDaemon.serverId,
+        workspaceId: secondary.workspaceId,
+      });
+
+      await test.step("workspace context opens the real wide panel bridge", async () => {
+        await switchWorkspaceViaSidebar({
+          page,
+          serverId: getServerId(),
+          workspaceId: primary.workspaceId,
+        });
+        await expectCommandAvailable(page, "Plugin global action");
+        await expectCommandAvailable(page, "Open plugin workspace");
+        await expectCommandUnavailable(page, "Open plugin agent");
+        await runCommand(page, "Open plugin workspace");
+        await expect(page.getByText(`Workspace bridge ${primary.workspaceId}`)).toBeVisible();
+        const renderCount = Number(
+          (await page.getByText(/Workspace renders \d+/).textContent())?.split(" ").at(-1),
+        );
+        await primaryClient.setWorkspaceTitle(primary.workspaceId, "Unrelated title update");
+        await expect(page.getByTestId("workspace-header-title")).toHaveText(
+          "Unrelated title update",
+        );
+        expect(
+          Number((await page.getByText(/Workspace renders \d+/).textContent())?.split(" ").at(-1)),
+        ).toBe(renderCount);
+        await expect(page.getByText(`Host ${getServerId()}`, { exact: true })).toBeVisible();
+        await expect(page.getByText("Layout wide", { exact: true })).toBeVisible();
+        await capture(page, testInfo, "plugin-workspace-panel-wide");
+      });
+
+      await test.step("direct and sidebar routes preserve same-id contribution kind", async () => {
+        await runCommand(page, "Open direct collision surface");
+        await expect(page.getByText("Direct collision surface", { exact: true })).toBeVisible();
+        await expect(page.getByText("Sidebar collision surface", { exact: true })).toHaveCount(0);
+        await page.getByTestId("plugin-surface-close").click();
+
+        await page.getByTestId(`plugin-sidebar-${PLUGIN_ID}-collision`).click();
+        await expect(page.getByText("Sidebar collision surface", { exact: true })).toBeVisible();
+        await expect(page.getByText("Direct collision surface", { exact: true })).toHaveCount(0);
+        await capture(page, testInfo, "plugin-surface-kind-collision");
+        await page.getByTestId("plugin-surface-close").click();
+      });
+
+      await test.step("switching hosts removes commands from an uninstalled host", async () => {
+        await switchWorkspaceViaSidebar({
+          page,
+          serverId: secondaryDaemon.serverId,
+          workspaceId: secondary.workspaceId,
+        });
+        await expectCommandUnavailable(page, "Plugin global action");
+        await expectCommandUnavailable(page, "Open plugin workspace");
+      });
+
+      await test.step("legacy sidebar URLs survive a full browser navigation", async () => {
+        await page.goto(`/h/${encodeURIComponent(getServerId())}/plugin/${PLUGIN_ID}/collision`);
+        await page.waitForURL(new RegExp(`/plugin/${PLUGIN_ID}/sidebar/collision(?:\\?.*)?$`));
+        await expect(page.getByText("Sidebar collision surface", { exact: true })).toBeVisible();
+      });
+
+      await test.step("agent context opens the compact panel with synchronous snapshots", async () => {
+        const agent = await primary.client.createAgent({
+          provider: "mock",
+          cwd: primary.repoPath,
+          workspaceId: primary.workspaceId,
+          title: "Plugin panel context agent",
+          model: "ten-second-stream",
+          modeId: "load-test",
+        });
+        await page.goto(buildAgentRoute(primary.workspaceId, agent.id));
+        await page.waitForURL(isSettledWorkspaceUrl, { timeout: 60_000 });
+        await page.setViewportSize(COMPACT_VIEWPORT);
+        await openCompactSidebar(page);
+        await runCommand(page, "Open plugin agent");
+        await closeMobileAgentSidebar(page);
+        await expectMobileAgentSidebarHidden(page);
+        await expect(page.getByText(`Agent bridge ${agent.id}`)).toBeVisible();
+        await expect(
+          page.getByText(`Workspace ${primary.workspaceId}`, { exact: true }),
+        ).toBeVisible();
+        await expect(page.getByText(`Host ${getServerId()}`, { exact: true })).toBeVisible();
+        await expect(page.getByText("Layout compact", { exact: true })).toBeVisible();
+        await capture(page, testInfo, "plugin-agent-panel-compact");
+      });
+
+      await test.step("removing the active plugin renders the panel unavailable", async () => {
+        await primaryClient.removePlugin(PLUGIN_ID);
+        await expect(
+          page.getByText("This plugin panel is unavailable.", { exact: true }),
+        ).toBeVisible({
+          timeout: 30_000,
+        });
+        await capture(page, testInfo, "plugin-panel-unavailable");
+        await openCompactSidebar(page);
+        await expectCommandUnavailable(page, "Open plugin agent");
+      });
+    } finally {
+      await primaryClient.removePlugin(PLUGIN_ID).catch(() => undefined);
+      await primaryClient
+        .patchDaemonConfig({ pluginsEnabled: previousConfig.config.pluginsEnabled ?? false })
+        .catch(() => undefined);
+      await primaryClient.close().catch(() => undefined);
+      await primary.cleanup().catch(() => undefined);
+      await secondary.cleanup().catch(() => undefined);
+      await secondaryDaemon.close().catch(() => undefined);
+      await rm(directory, { recursive: true, force: true });
+    }
+  });
+});

packages/app/e2e/browser/projects-settings.spec.ts

@@ -9,12 +9,14 @@ import {
   clickReloadProjectSettings,
   clickRetryProjectSettingsSave,
   clickSaveProjectSettings,
+  commitPaseoConfig,
   corruptPaseoConfig,
   editWorktreeSetup,
   expectEmptyScriptList,
   expectProjectHostContextHidden,
   expectNoEditableTarget,
   expectNoProjectSettingsError,
+  expectNoUncommittedSetupWarning,
   expectProjectEditFailed,
   expectProjectEditName,
   expectProjectEditSaved,
@@ -27,6 +29,7 @@ import {
   expectSaveButtonDisabled,
   expectScriptRowCount,
   expectWriteFailedCalloutActions,
+  expectUncommittedSetupWarning,
   fillProjectIconUrl,
   fillProjectName,
   installDaemonConnectionGate,
@@ -42,12 +45,19 @@ import {
   unblockPaseoConfigWrites,
 } from "../support/helpers/project-settings";
 import { gotoAppShell } from "../support/helpers/app";
+import { openCompactSettings } from "../support/helpers/settings";
 import {
   addProjectFlowInput,
   chooseAddProjectMethod,
   openAddProjectFlow,
 } from "../support/helpers/add-project-flow";
 import { createTempGitRepo } from "../support/helpers/workspace";
+import {
+  buildOpenProjectRoute,
+  buildProjectsSettingsRoute,
+  buildSettingsRoute,
+} from "@/utils/host-routes";
+import { getServerId } from "../support/helpers/server-id";
 
 const updatedSetup = ["npm install", "npm run build"];
 
@@ -216,9 +226,16 @@ test.describe("Projects settings", () => {
   test("user edits worktree setup from the projects page", async ({ page, editableProject }) => {
     await openProjects(page);
     await openProjectSettings(page, editableProject.name);
+    await expectNoUncommittedSetupWarning(page);
     await editWorktreeSetup(page, updatedSetup);
     await clickSaveProjectSettings(page);
     await expectProjectConfigSaved(editableProject);
+    await expectUncommittedSetupWarning(page);
+
+    commitPaseoConfig(editableProject.path);
+    await returnToProjectsList(page);
+    await openProjectSettings(page, editableProject.name);
+    await expectNoUncommittedSetupWarning(page);
   });
 
   test("project navigation stays inside the selected host", async ({ page, editableProject }) => {
@@ -311,6 +328,33 @@ test.describe("Projects settings", () => {
   });
 });
 
+test.describe("Projects settings — compact navigation", () => {
+  test.use({ viewport: { width: 390, height: 844 } });
+
+  test("project Back returns through Projects and Settings without reopening the project", async ({
+    page,
+    editableProject,
+  }) => {
+    await gotoAppShell(page);
+    await openCompactSettings(page, buildOpenProjectRoute());
+    await page.getByRole("button", { name: "Projects", exact: true }).click();
+    await expect(page).toHaveURL(buildProjectsSettingsRoute(getServerId()));
+
+    await openProjectSettings(page, editableProject.name);
+    await expect(page.getByRole("button", { name: "Back", exact: true })).toHaveCount(1);
+    await expect(page.getByRole("button", { name: "Back to projects", exact: true })).toHaveCount(
+      0,
+    );
+
+    await page.getByRole("button", { name: "Back", exact: true }).click();
+    await expect(page).toHaveURL(buildProjectsSettingsRoute(getServerId()));
+
+    await page.getByRole("button", { name: "Back", exact: true }).click();
+    await expect(page).toHaveURL(buildSettingsRoute());
+    await expect(page.getByTestId("settings-sidebar")).toBeVisible();
+  });
+});
+
 test.describe("Projects settings — error UX", () => {
   test("stale-write callout appears on save, disables save, and reload clears it", async ({
     page,
@@ -439,5 +483,7 @@ test.describe("Projects settings — error UX", () => {
 
     await expectScriptRowCount(page, 0);
     await expectEmptyScriptList(page);
+    await clickSaveProjectSettings(page);
+    await expectNoUncommittedSetupWarning(page);
   });
 });

packages/app/e2e/browser/provider-settings-refresh.spec.ts

@@ -21,21 +21,22 @@ async function openMockAgentAtMobileBreakpoint(page: Page) {
 
 async function openProviderSettingsFromModelSelector(page: Page) {
   await page.getByRole("button", { name: /Select model/ }).click();
-  await expect(page.getByLabel("Bottom Sheet", { exact: true })).toBeVisible({ timeout: 10_000 });
+  const configuration = page.getByTestId("agent-controls-model-sheet");
+  await expect(configuration).toBeVisible({ timeout: 10_000 });
+  await page.getByTestId("agent-controls-model").click();
 
-  const openCodeRow = page.getByText("OpenCode", { exact: true }).first();
-  if (await openCodeRow.isVisible().catch(() => false)) {
-    await openCodeRow.click();
-  }
+  const modelBrowser = page.getByTestId("agent-controls-model-browser-sheet");
+  await expect(modelBrowser).toBeVisible({ timeout: 10_000 });
 
   await page.getByRole("button", { name: /Open .* settings/ }).click();
   await expect(page.getByTestId("provider-settings-sheet")).toBeVisible({ timeout: 10_000 });
 }
 
-async function expectModelSelectorVisible(page: Page) {
-  await expect(page.getByRole("button", { name: /Open .* settings/ })).toBeVisible({
+async function expectModelBrowserVisible(page: Page) {
+  await expect(page.getByTestId("agent-controls-model-browser-sheet")).toBeVisible({
     timeout: 10_000,
   });
+  await expect(page.getByRole("button", { name: /Open .* settings/ })).toBeVisible();
 }
 
 async function closeTopSheet(page: Page) {
@@ -160,7 +161,6 @@ test.describe("provider settings overlay stack", () => {
       await commandCenter.getByText("Add project", { exact: true }).click();
       const addProject = page.getByTestId("add-project-flow");
       await expect(addProject).toBeVisible({ timeout: 10_000 });
-      await expect(addProject.getByTestId("add-project-flow-input")).toBeFocused();
       await page.keyboard.press("Escape");
       await expect(addProject).not.toBeVisible({ timeout: 10_000 });
       await expect(selector).toBeVisible();
@@ -181,7 +181,7 @@ test.describe("provider settings overlay stack", () => {
     }
   });
 
-  test("provider settings and children close back through the model selector stack", async ({
+  test("provider settings and children close back through the model browser to configuration", async ({
     page,
   }) => {
     test.setTimeout(180_000);
@@ -193,13 +193,18 @@ test.describe("provider settings overlay stack", () => {
       await exerciseProviderSettingsStack(page);
       await closeSheetByHeaderButton(page, "provider-settings-sheet");
 
-      await expectModelSelectorVisible(page);
+      await expectModelBrowserVisible(page);
       await page.getByRole("button", { name: /Open .* settings/ }).click();
       await expect(page.getByTestId("provider-settings-sheet")).toBeVisible({ timeout: 10_000 });
       await exerciseProviderSettingsStack(page);
       await closeSheetByHeaderButton(page, "provider-settings-sheet");
 
-      await expectModelSelectorVisible(page);
+      await expectModelBrowserVisible(page);
+      await closeTopSheet(page);
+      await expect(page.getByTestId("agent-controls-model-browser-sheet")).not.toBeVisible({
+        timeout: 10_000,
+      });
+      await expect(page.getByTestId("agent-controls-settings-list")).toBeVisible();
       await closeTopSheet(page);
     } finally {
       await session.cleanup();