A Bash call with run_in_background and an Agent call return at once
('Command running in background…' / 'Async agent launched…'), so the
chat's task list counted them finished immediately and showed 0 tasks.
They now stay running until Claude Code's task notification (the
<task-notification> text or the SDK's task_notification message)
reports completion, with the exit code from its summary.
Persist three global ntfy switches in server_settings: run_finished
(default on: auto-mode runs notify on success, read inside the
transition transaction; on/off modes unchanged), chat_needs_you
(default on: gates chat question/approval alerts) and
chat_reply_finished (default off: one alert per finished assistant
turn). Expose notification_settings / set_notification_settings over
the websocket protocol, add a Phone notifications section with
Material and Cupertino checkboxes in Settings, document the settings,
and keep the cli, tui, stage and pipeline test packages from
inheriting a real ntfy configuration.
Operator-attached images render as rounded thumbnails above the user
bubble and open a zoomable Material/Cupertino viewer. Inline data: URIs
decode locally; file:// attachment-store URIs load through the new
chat_attachment_image websocket request, which serves only regular image
files directly inside the private attachment store (os.Root, <= 4 MiB,
sniffed raster types). Missing files show an 'Image no longer available'
placeholder; other files keep the file row.
Run notifications now always click through to pipelines://run/<id>; a valid
PIPELINES_WEB_PUBLIC_URL adds an "Open in browser" ntfy view action instead
of replacing the click. Queued rows derive links from the run id at delivery,
so legacy rows holding a web URL still open the app.
Unresolved chat questions/forms and permission requests publish one async,
bounded notification to the system destination (pipelines://chat/<id> plus the
browser action), deduplicated per form/request id across reconnects and
replays. stage.PublishNtfy gains optional, strictly validated view actions.
Server tests now clear the operator's real ntfy environment.
Six computed Flutter runs stopped at the export handoff gate because the
rendered session.html cap was 2 MiB while their complete sessions (well inside
the 8 MiB source read bound) rendered between 2.25 and 4.28 MiB once html
escaping and the readable projection duplicated native text.
Raise SessionExportFileLimit to 8 MiB, the server's existing per-artifact file
bound, so a complete export is still listed and served. Normal export
verification, versioned repair and the handoff gate keep using the single
shared constant. The source read bound is unchanged, and an over-cap render is
still visibly truncated and recorded partial.
Add focused regressions for a complete export above the old cap, an over-new-cap
file rejected as partial, repair above the old cap leaving original evidence
untouched, the gate accepting the larger repair, and the constant staying inside
the artifact file bound. Update the artifact, handoff-gate and operations docs
and record the diagnosis in docs/notes/session-export-render-cap.md.
Read a bounded newest-first page of eight assistants (4 MiB, no cursor traversal), selecting context and completed generation rate independently from the latest valid measurement matching Session.Info.model provider/id. Require assistant model identity; never pair an old model's measurements with the selected model's window. Context remains unknown across a newer compaction until a subsequent prompt reports usage; the last completed matching rate may survive compaction.
Unknown model metadata is cached for 15 seconds instead of five minutes, while known capacities retain their five-minute cache. Provenance remains canonical V2 Model.Ref, TokenUsage and request-created/streamed/completed fields cited in chat_usage.go. No live endpoint access, inference or service changes.
Isolated HTTP regressions cover in-flight/null/empty newest tokens with prior known metrics, post-compaction boundaries, independent prompt/rate selection, switched/missing model identity, duplicate known-zero measurements, over-limit page rejection and metadata recovery. Checked full internal/opencode suite, focused usage race suite, go vet ./internal/opencode and git diff --check.
Read V2 session accounting, latest assistant prompt/cache components and native request-to-stream boundary. Resolve capacity by location/provider/catalog id with bounded metadata caching. Coalesce active/final readbacks; retain message-less session snapshots without changing admission or compaction state.
Provenance: https://opencode.ai/v2/openapi.json and anomalyco/opencode V2 4617210822bdbb31e5749751afbfb4584ccedac8: packages/schema/src/token-usage.ts, session-event.ts; packages/core/src/session/usage.ts, projector.ts, message-updater.ts, runner/step.ts. Native output is visible output, reasoning separate. Session totals include all five non-overlapping components. Generation rate uses created-to-streamed, not tool-inclusive completion.
Limitations: metrics remain unknown for missing components, absent timing/model metadata, reads beyond bounded latest-message/catalog budgets, or context preceding compaction. No local pricing, token estimation or default capacity. V1 gets no invented cumulative metrics; older V2 events remain available when readback is unavailable. No deployment or protocol/web/docs changes.
Checks: go test ./...; focused usage/stream race tests; go vet ./internal/opencode ./internal/server; go build ./...; git diff --check.
Verify typed ThreadTokenUsage/TokenUsageBreakdown/Thread/Turn against openai/codex@6fb1c3903d. Use native totalTokens and latest-call context tokens without re-adding cached/reasoning subsets. Replay bounded per-connection snapshots and clear on room/source release.
Cost stays unavailable: no subscription USD usage is reported by this notification. Throughput stays unavailable even with native Turn.durationMs because neither last-call nor cumulative-session output is a compatible turn-specific numerator. Thread readback exposes no token usage; no extra reads, resume, or live provider calls added.
Checks: go test -race ./internal/codex; go vet ./internal/codex; go test ./internal/server ./internal/protocol; git diff --check.
Select native installs only via explicit PIPELINES_PREFIX or the matching owned prefix/bin/pipelines executable. Route enable/status/disable to the existing native supervisor; validate registration ownership, label and command, keep preview read-only, and disable login registration before stopping. Never fall back to a standalone service for an invalid explicit prefix.
Fresh native installs explicitly enable authoring, while reinstalls preserve false or missing legacy role settings. Add explicit --allow-authoring/--no-authoring overrides and pass the server flag only for manifest true. Print a copyable native enable command and update installation help/docs. No Docker configuration changes.
Checks: 54 mocked native installer/lifecycle tests; 8 entrypoint tests; offline Go installation/CLI tests and vet; gofmt clean; shell syntax and diff checks. No live install, restart, deployment, permission/config changes, or native Linux verification.