fix(ci): pin codex CLI to 0.105.0 and improve turn_failed diagnostics

Playwright archive-tab tests fail because CI installs @openai/codex@latest
(0.120.0) which has breaking protocol changes vs the known-working 0.105.0.
Pin the version and add diagnostics for future debugging: elevate turn_failed
log from TRACE to WARN, and include error details in test assertions.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.github/workflows/ci.yml

@@ -120,7 +120,7 @@ jobs:
         run: npm run build --workspace=@getpaseo/server
 
       - name: Install agent CLIs for provider tests
-        run: npm install -g @openai/codex opencode-ai
+        run: npm install -g @openai/codex@0.105.0 opencode-ai
 
       - name: Run Playwright E2E tests
         run: npm run test:e2e --workspace=@getpaseo/app
@@ -170,7 +170,7 @@ jobs:
         run: npm install
 
       - name: Install agent CLIs for provider tests
-        run: npm install -g @openai/codex opencode-ai
+        run: npm install -g @openai/codex@0.105.0 opencode-ai
 
       - name: Build highlight dependency
         run: npm run build --workspace=@getpaseo/highlight

packages/app/e2e/helpers/archive-tab.ts

@@ -118,7 +118,7 @@ export async function createIdleAgent(
   });
   const finished = await client.waitForFinish(created.id, 120_000);
   if (finished.status !== "idle") {
-    throw new Error(`Expected agent ${created.id} to become idle, got ${finished.status}.`);
+    throw new Error(`Expected agent ${created.id} to become idle, got ${finished.status}. Error: ${JSON.stringify((finished as Record<string, unknown>).error ?? "unknown")}`);
   }
   return {
     id: created.id,

packages/server/src/server/agent/agent-manager.ts

@@ -2180,7 +2180,7 @@ export class AgentManager {
         void this.refreshRuntimeInfo(agent);
         break;
       case "turn_failed":
-        this.logger.trace(
+        this.logger.warn(
           {
             agentId: agent.id,
             lifecycle: agent.lifecycle,