[kernel-1116] browser events: add external events

[archandatta]

As per plan -> #217

---

Note

Medium Risk
Adds new public API endpoints that inject and stream capture-session events, plus changes capture-session stop/publish behavior to emit terminal/system events; incorrect validation or streaming behavior could impact clients and event pipeline reliability.

Overview
Adds an external events API: POST /events/publish to inject validated caller-supplied events into the active capture session (with reserved types blocked and source.kind forced), and GET /events/stream to stream envelopes over SSE with Last-Event-ID resume support, keepalives, and explicit events_dropped/session_ended signaling.

Updates the events pipeline to support this: CaptureSession now exposes Active(), publishes a synthetic session_ended event on Stop(), and adds PublishUnfiltered() to bypass category filtering for externally-initiated events. OpenAPI/generated client/server code is regenerated for the new endpoints and enums, and SSE responses now include Cache-Control: no-cache and X-Accel-Buffering: no headers; tests are updated/added to cover the new event lifecycle and enum rename.

^{Reviewed by Cursor Bugbot for commit 4cb8f0e. Bugbot is set up for automated code reviews on this repo. Configure here.}

[firetiger-agent]

Firetiger deploy monitoring skipped

This PR didn't match the auto-monitor filter configured on your GitHub connection:

Any PR that changes the kernel API. Monitor changes to API endpoints (packages/api/cmd/api/) and Temporal workflows (packages/api/lib/temporal) in the kernel repo

Reason: PR title and branch suggest browser events feature work, but no file changes are provided to confirm modifications to API endpoints or Temporal workflows.

To monitor this PR anyway, reply with @firetiger monitor this.

[Sayan-]

matches plan well and deviations from it are reasonable!

[Sayan-]

should seq be monotonic across start/stop cycles? If so, I think Start shouldn't reset it here. That would make seq represent the process-wide event stream position rather than the position within a single capture session.

[archandatta]

I see your point, here are things to consider with how this works currently

• every envelope already carries capture_session_id, so i think (session_id, seq) would be sufficient
• SSE clients don't cross session boundaries, session_ended terminates the stream, so not sure if monotonic seq applies fully

the fix would also require a non-trivial ring buffer refactor to avoid dropped records

[Sayan-]

could we add one lifecycle test for the new event flow? Something like start capture session, publish an event through PublishEvent, read it from StreamEvents, then stop and verify the stream receives session_ended.

[archandatta]

4cb8f0e

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 4cb8f0e. Configure here.}

[cursor]

Stop response seq off-by-one after session_ended publish

Medium Severity

StopCaptureSession calls buildSessionResponse() (which reads Seq()) before s.captureSession.Stop(). Now that Stop() publishes a synthetic session_ended event via publishLocked — which increments s.seq — the Seq in the returned response is one less than the actual final sequence number. Clients comparing the stop response's Seq with the session_ended envelope's seq on the SSE stream will see an inconsistency.

 

^{Reviewed by Cursor Bugbot for commit 4cb8f0e. Configure here.}