Browser tools: full accessibility snapshot, trusted input, dialogs, evaluate, and tab controls (#1881)

* feat(browser): replace flat snapshot with ARIA tree and page-side refs

* feat(browser): trusted CDP input with actionability waits

* feat(browser): auto-handle javascript dialogs

* feat(browser): add browser_evaluate

* feat(browser): add scroll, resize, and close_tab tools

* fix(browser): review fixes for snapshot/input/dialog seams

* fix(browser): address review findings on dialog capture, truncation, and keypress targets

* fix(harness): drop resize check that cannot run under parked compositor surfaces

* fix(browser): second review round — snapshot fidelity, key input, dialog scoping
This commit is contained in:
Mohamed Boudra
2026-07-04 01:38:07 +02:00
committed by GitHub
parent b28f291493
commit 9c220c2451
29 changed files with 5653 additions and 949 deletions

View File

@@ -21,6 +21,18 @@ Run it with the repo Electron:
npm run capture-harness --workspace=@getpaseo/desktop
```
Run the browser automation fixture with:
```bash
PASEO_CAPTURE_HARNESS_GROUP=automation npm run capture-harness --workspace=@getpaseo/desktop
```
The automation group uses a real guest webview to verify the page-side ref contract:
ARIA-like snapshot text includes headings, static text, and controls; refs survive
`pushState` when the element still matches; same-URL rerenders stale old refs; and a
file-input ref can be resolved to a CDP backend node id for upload. It also verifies
page-context evaluation, including passing a resolved ref element as the function argument.
On macOS the harness process must set `app.setActivationPolicy("accessory")` and
hide the Dock icon before creating any window. `showInactive()` only prevents window
focus; a normal Electron app launch can still activate the app and steal focus.