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.
Raise phone body/meta type, controls and dock targets to 44/56px, make the chat
back arrow a visible 44px control, and give the star/menu 44px hit areas without
covering the chat status. Keep the shipped skill free of personal device and
project capture paths.
- Persist chat stars in the operator-session registry (chat_star) so every
browser/device shows the same Starred group; migrate old local stars once.
- Add a compact pipeline graph shape to list responses and draw a lightweight
list miniature; replace the cryptic id/version line with the stage count.
- Surface OpenCode form questions as a chat card (chat_answer), replay a
pending form to late subscribers, and notify for background questions.
- Make the mobile dock's Search cell equal to every destination, keep the
tab row visible with 100svh, and let Source use the full pane height.
- Let a fresh sign-in auto-connect to its own origin; use a Lax session cookie
with Expires so Remember me survives app/link launches.
- Keep shipped starters and the skill agnostic: no repository paths, project
areas or personal device commands; add Diagnose/Fix and Research starters.
- Refresh the installed skill and document the new browser conventions.
Each model visit now gets <run>/stages/<seq>-<stage>/ with input.md,
memory.md, output.md, private/ and a shared run-level public/. The model
reads the two input files, follows the stage's instructions verbatim and
returns complete markdown; only the # outcome section is parsed.
- add internal/turn layout, render and output parsing
- write input.md/memory.md before a visit and output.md from the final
message; inventory all turn files as stage artifacts
- accumulate completed turns and system-written public file additions in
memory, including across retries and resume
- remove the JSON envelope, strict field validation, line format and the
repair ladder for model stages
- republish pipeline authoring prompts and docs in the new style
Checks: go build ./..., go vet ./..., go test -count=1 ./..., gofmt -l .
Drop the resource allocation feature end to end: store tables and migration, runner scheduling and cleanup, protocol types and commands, CLI commands, TUI screens, tests and docs. Existing databases open cleanly and definitions carrying a retired resources block still validate.
A stage's session was exported twice, as session.md and session.html, and both
files were presented. It is now one html file; the markdown export is retired.
- internal/stage/session_export.go: ExportSession renders and writes exactly one
file, session.html, via writeSessionFile; renderSessionMarkdown and its helpers
are gone; SessionExportResult carries HTML/HTMLBytes only. The source read, the
8 MiB source / 2 MiB rendered bounds, the truncation banner, the
never-overwrite rule and the honest empty-session handling are unchanged.
- internal/protocol/runs.go: SessionExport drops markdown/markdown_bytes (a
record written before the change still decodes; unknown fields are ignored),
and adds LegacySessionMarkdownArtifact with RunStep.PresentedArtifacts(), the
one Go rule that hides the retired session.md from every artifact listing.
- internal/server/runner.go: evidence.Markdown/MarkdownBytes are no longer copied.
- internal/tui: the artifacts panel, the recorded count and the per-visit summary
derive from PresentedArtifacts(), so the terminal never lists session.md.
- web/src/app.js: the spike's renderArtifacts applies the same filter.
- android: the Session section is the html export alone in a WebView — SessionView,
the Markdown/HTML toggle and the session markdown path are removed;
presentedArtifacts(step) filters the retired session.md out of the per-visit
gallery, the run gallery and its count, the verdict/selection lists and the
timeline indicator; data/protocol SessionExport drops the markdown fields.
- Tests pin the single html file, prove no session.md is written and that a
pre-existing one is neither overwritten nor presented, and prove a legacy
export record still loads and its session.md is hidden.
The legacy session.md of an older run stays on disk (dated evidence); it is only
no longer produced and no longer presented.
Checks (macOS, Go 1.24, Android Gradle 8.11.1, 18 Oct 2026): gofmt -l . prints
nothing; go build ./... && go vet ./... clean; go test ./internal/... ./cmd/...
-> 19 packages ok, 0 failures; node --check web/src/app.js ok and node --test
web/test -> 20/20; gradle --offline --no-daemon :app:assembleDebug
:app:testDebugUnitTest -> BUILD SUCCESSFUL.
Packaged references checked: internal/author/skill/references/delivery.md and
operations.md both described the session.md+session.html pair, so both are
updated here to describe the single session.html export and the retired, hidden
session.md. README.md and internal/author/skill/SKILL.md carry no session.md
reference and install.go's managed inventory is unchanged (no reference added or
removed). docs/notes/session-html-only.md records the remaining stale shared
documentation (docs/stage-artifacts.md, docs/app-integration-verification.md) for
the docs-fold writer; shared docs are not edited by this item.
Rewrite the workflow documentation to describe how the system actually works:
the agent (the session) is the brain and advances the queue by hand through the
`pipelines` CLI. There is no orchestration script, no driver, no polling loop and
no lane machinery, and none may be written.
- references/work.md: rewritten end to end. The queue-driver script
(drivers/workd.sh), the claim-file ledger with atomic renames, the PAUSED
marker and the poll loop are gone. State is now the server (authoritative for
run state), the queue file (authoritative for intent) and the conversation.
The agent reads the queue, picks the item, creates its worktree/branch and
launches it with `pipelines run <pipeline> <brief> --json --yes`; progress is
event-driven via a background `pipelines wait <run-id> [...]` (never polled);
on completion the agent reconciles with `pipelines inspect <run-id> --json`,
verifies it landed and starts the next item.
- references/queue.md: the landing gate runs the command directly
(`git merge-base --is-ancestor <branch> main`, exit 0 = landed) instead of a
`land-check.sh` helper; the never-delete-branch / never-remove-worktree rules
are kept, and the waiting section states the agent advances the queue by hand.
- references/orchestration.md: intro and "Running and watching" state the
agent-driven, event-driven model; "Pipeline choice" names the current set
(Android, TUI, Server, Generic, Plan and land, UI sweep, Probe) by area and
records the retired names.
- references/debugging.md: "Duplicated work after a driver restart" became
"…after a session restart", resuming from the queue file not a cursor file.
- SKILL.md: the routing row no longer says "queue driver".
- AGENTS.md: the maintenance workflow names the current pipeline set instead of
the retired `Pipelines maintenance - plan implement review`.
- internal/tui/AGENTS.md: the restart line no longer names lanes.
- docs/notes/agent-driven-work-model.md: the docs-fold note recording that
docs/cli.md needs its "Scripted runs"/`--wait` framing aligned to the
event-driven model, and that the dated evidence docs stay as they are.
Checked before committing: no packaged template directory exists in this repo, so
none needed updating; internal/author/install.go's managed file inventory needs no
change (no packaged reference was added or removed — the 13 listed files all
match internal/author/skill/references/); no stale `queue driver`/`workd.sh`/
`claim-file` references remain anywhere under internal/author. Genuine device
tooling (android/scripts/capture_png.sh-style) and server-side polling prose were
left alone.
Checks: `go test ./internal/author -count=1` ok; `go build ./...` clean;
`go vet ./...` clean; `gofmt -l .` clean.
The lane machinery is gone: work is a markdown queue, items run in parallel in
their own worktrees, the committer stage lands its own work, land-check.sh gates
cleanup, branches are never deleted, and pipelines wait (rather than a driver
script) is what wakes a session. references/lanes.md becomes references/queue.md;
the remaining lane references in work/orchestration/debugging/delivery/format and
SKILL.md are rewritten, and the installer now prunes a managed file it no longer
bundles when the copy is unmodified.
Every item wrote the shared, append-only docs (android/README.md, docs/cli.md and
the packaged skill references), so every merge conflicted on prose. Keep the duty
to notice stale docs; move the writing to one deliberate place.
- Per-item note, not a shared edit. An item that makes a shared document stale
writes docs/notes/<slug>.md - unique per item, so two items never touch the same
file - and commits that note with its change. It never edits a README, topic doc
or packaged skill reference itself.
- One writer. The manual-only docs-fold pipeline, bundled as the fifth starter
examples/docs-fold.md, folds every note under docs/notes/ into the shared
document it names and deletes the consumed notes in its own commit; an empty
notes directory is a success no-op. Operator trigger:
pipelines new "Docs fold" --template docs-fold --no-open; pipelines submit
<draft.md>; pipelines run "Docs fold".
- Implemented where the convention lives: orchestration.md (brief item 5, the
hard-constraint bullet, a "Single documentation writer" section and the
"Pipeline choice" committer bullet), the canonical committer checklist in
format.md (new step 6), the committer prompts in examples/task.md and
examples/three-stage.md, the new examples/docs-fold.md with its catalog and
README entries, and the repository guideline AGENTS.md. Test-expectation counts
(examples/catalog_test.go, internal/server/templates_test.go,
scripts/smoke-first-install.py) follow the fifth starter.
Reflected where it belongs: the packaged templates (examples/task.md,
examples/three-stage.md, the new examples/docs-fold.md), the packaged skill
references (orchestration.md, format.md) and the repo guideline AGENTS.md all carry
the new convention. No new reference file was added, so install.go's managed
inventory is unchanged. The shared descriptive docs were deliberately not
rewritten: this change makes README.md:191, docs/cli.md:170 and
internal/author/skill/references/operations.md:13 name four starters where there are
now five. Those stay as they are - correcting them now is exactly the per-item
shared edit this change removes - and the stale-doc note
docs/notes/docs-fold-single-writer.md records each one for the docs-fold writer to
fold. Nothing else needed updating. No new dependency; no Android, Gradle, server,
installer, Docker or published-definition change; no definition published over the
running pipelines.
Evidence: go test -count=1 ./... passed; go build ./... and go vet on the touched
packages clean; ./internal/author/... covers embedding, reference links, the managed
inventory and install/upgrade/protection; ./examples/... validates the fifth starter
and its manual-only, committer-terminated graph; ./internal/server/... covers the
served catalog. Not run: the docs-fold writer against the shared docs (nothing to
fold yet); no install or deploy.
(cherry picked from commit f29656cf084d7eee2e033f2a37a9d231d7acaaf0)
Carry the usage a model stage already measures for its budget
(stage.Budget.Snapshot()) onto the step as typed evidence and total it in
one pass, rather than a second harness read, a transcript parse or an
output-length estimate. A shared budget across stage revisits records only
the visit's increment, and a revisit that observes no new report records no
usage instead of a fabricated known zero.
- protocol: StageUsage/RunUsage/TotalUsage and UsageTotals.
- pipeline: step_usage table + migration, COALESCEd into the one step
projection, FinishStep persistence, one json_extract history aggregate,
retention removes the rows with their run.
- server: stageUsage maps the budget snapshot, stamping per-stage model and
usage for every model stage (including failures).
- cli: inspect/stats human and JSON output.
- budget/pi: retain the reported in/out split, mark a total-only (Pi) report
split-unknown, and keep an unreported cost unknown (never zero).
Evidence and docs updated in the packaged skill references and docs/cli.md;
the live database has no step_usage rows yet (all history predates this), so
per-run figures are reported as unavailable, not zero.
Add an opt-in way to broker chat from an operator-supplied OpenCode service
instead of the instance the server spawns for itself, and a way to adopt a
session that already exists on that endpoint so it appears in chat_sessions and
can be continued from the app.
- Endpoint selection: the `-opencode-endpoint` server flag (defaulting to
`PIPELINES_OPENCODE_ENDPOINT`) opts in; precedence is explicit flag >
explicit environment > the endpoint recorded on a previous run, with the
spawned instance as the default. opencode.OpenEndpoint validates the
reference once (http(s), host, no userinfo/query/fragment/path), detects the
backend from its own /api/info (v2) or /global/health (v1) contract, and
requires a streamable text/event-stream event path. It owns no process
(Close is a no-op) and never calls startServe; every failure names the
endpoint and disables chat rather than silently falling back. An
authenticated endpoint takes an optional password only through the private
file PIPELINES_OPENCODE_ENDPOINT_PASSWORD_FILE, with
PIPELINES_OPENCODE_ENDPOINT_USERNAME overriding the detected default.
- Persistence: the validated choice is recorded in an additive single-row
server_settings table so a restart with no flag/environment reuses it; an
unreachable recorded endpoint disables chat with a clear message and no
fallback.
- chat_adopt: a new protocol request (ChatAdoptRequest) with handleChatAdopt in
internal/server/chat.go. It refuses a nil broker/registry/payload and a
blank or foreign-shaped id before any endpoint call, reads the endpoint via
the shared Chat.Session (which reuses Sessions/decodeChatSessions), records
through the existing RecordOperatorSession with the endpoint's directory and
title verbatim, and echoes the adopted session plus its registry row like
chat_create. A missing id is refused naming the id and records nothing; an
endpoint record with an empty directory is refused rather than substituting
the server's working directory or home.
- CLI: `pipelines chat-adopt <session-id> [--json]` (usage, notes, help
inventory, dispatch) with docs/cli.md updated under Chat sessions.
- Autostart preserves the endpoint variables (and the password-file reference,
never an inline secret) via serviceEnvironmentKeys.
Checks before committing:
- Package location: the change is user-facing chat behaviour, so the packaged
authoring skill reference internal/author/skill/references/operations.md is
updated in the same change (Brokered OpenCode chat section) with a dated
verification-evidence entry and the not-deployed state. No new reference file
is added, so internal/author/install.go's managed inventory is unchanged.
- Packaged templates: this repository packages no template set for this
behaviour and none references the chat endpoint, so no template needs
updating.
- Verification: `go build ./...`, `go vet` on the touched packages, `gofmt -l`
clean and `go test ./...` all pass. The unit tests cover the endpoint
validation/precedence, the adopt happy path and refusals, the recorded
endpoint round-trip, the protocol round-trip and the CLI dispatch. The live
scratch-server check against the operator's real endpoint was not run in this
change.
Nothing unrelated is staged; docs left stale (README and other shared documents)
are not edited here and are for a single folding writer.
Every OpenCode-harness model stage now writes its real session into its own
step artifacts directory as session.md and session.html, so the existing
bounded artifact listing, the /artifacts/<run>/<step>/<name> serving path and
the artifact scan find and serve them with no special casing, while the stage
keeps its bounded preview in the run record unchanged.
- internal/opencode/export.go adds the one shared native export invocation and
the exact session-ID/messages validation used by both Connection.Messages and
the new ExportSession, so the readback path does not diverge. The backend's
published subcommand is used (v1 `export <id>`, v2 `session export <id>`),
the source read is capped at 8 MiB, and a larger session is read as a prefix
and reported truncated instead of failing. boundedOutput.ReadFrom keeps the
cap effective so overflow and the truncation path are real.
- internal/stage/session_export.go renders markdown and html with a metadata
header, every role and every part (unknown part types as raw JSON, every
session text escaped in html), atomic temp+rename writes that never overwrite
an existing file, a visible truncation banner at both ends when the source or
the 2 MiB rendered cap cuts it, and an honest 0-message header for an empty
session rather than fabricated text.
- protocol.SessionExport and RunStep.SessionExport are stored in a new
step_session_exports table (CREATE TABLE IF NOT EXISTS, so an old database
gains an empty table with no migration) loaded through the existing main
SELECT alongside the other per-step evidence and deleted by run retention.
- internal/server/runner.go exports after the post-stage filediff capture and
before scanArtifacts, with context.WithTimeout(context.WithoutCancel(ctx),
10s); it never waits for session idleness and a failure or a Pi stage (no
native export path) records a sanitized diagnostic and continues, never
changing the stage status, outcome, target, timing or result validation.
Reflected where it belongs: the packaged skill references
(internal/author/skill/references/operations.md and delivery.md) carry the new
Session export rules and the run-artifacts bullet, and docs/stage-artifacts.md
documents the names, the <run>/<step> paths, the served URL, the caps, the
truncation labels, failure isolation and the Pi limitation. The packaged
templates (examples/*.md) only tell a stage to save its own artifacts and say
nothing about the server-written session export, so none needed updating; no
new reference file was added, so install.go's managed inventory is unchanged.
Shared documentation: README.md's per-step artifacts sentence is now stale - it
should name the exported session alongside images, screenshots, reports and
logs. It is deliberately not committed here; the shared documents are folded by
the single documentation writer.
Checks: gofmt -l on the touched packages produced no output; go vet ./...
passed; go test -race -count=1 ./... passed.
Keeps this branch's advisory review edge, bounded blocked_retryable implementer
edge and their tests alongside main's run-attention/ntfy notification work, the
split of the Android chat conversation screen, and the retention/history changes.
No conflicts: the two sides touch disjoint files.
Cut the review loop's repeat cost and make a blocked stage actionable, at the
definition level only, with the runner's existing visit counting and caps.
The published maintenance, quick and focused pipelines gain a reviewer
`advisory -> next stage` edge for non-blocking notes, so an advisory-only review
continues the run instead of buying another implementer visit; `rejected ->
implementer` and its unchanged {"max_visits":3,"on":["rejected"]} limit keep a
must-fix finding looping and still terminate. The implementer gains
blocked_retryable -> implementer with {"max_visits":5,"on":["blocked_retryable"]}
for a transient environment/tooling/device problem a fresh attempt could
survive, while blocked -> #failure stays terminal. The cap is at least the
reviewer's rejection budget so the implementer's own pre-execution cap cannot
preempt the reviewer's rejected control.
No production Go code changed. New tests in internal/server/review_contract_test.go
cover advisory-only never revisiting the implementer, must-fix still looping to
the reviewer limit, the implementer cap not preempting the reviewer, retryable
blocks bounded then failing closed, a retry that recovers, and a terminal block
failing immediately; concurrency/control/human tests share the new submitRunner
and startRun seams.
Documentation checked: internal/author/skill/references/control.md and
docs/harnesses-and-control.md carry the new edges and the visit-cap rule;
no packaged template defines these pipelines (they are published immutable
versions), and docs/model-results.md needs no change because the {outcome,output}
result contract is unchanged. README and other shared docs deliberately untouched.
Verified: go build ./... and go test -count=1 ./... pass.
Residual limitation: the must-fix-vs-advisory split is carried by the prompt
contract, not a machine-readable marker, because the recorded findings have none;
a model that still returns rejected for advisory-only notes would still buy a
visit. All 18 recorded implementer blocked runs classify terminal, so the
blocked_retryable edge recovers 0 of them.
Nothing told the operator that a run was waiting for a person or had failed; they
found out by going looking. A run that needs attention now reaches the ntfy they
already run, and the app answers "what needs me now" in one live, ordered surface.
Server (smallest change, reusing the existing publisher):
- internal/server/notify.go adds notifyRun, called from finishRunState (failed)
and when a run parks on a human (runner.go runLoop). It publishes to the same
already-configured `system` destination through the unchanged stage.PublishNtfy,
so there is no second publisher, topic or service and the token handling is the
stage's own. runAttentionEvent maps only waiting and failed; queued, running,
succeeded, canceled and interrupted publish nothing.
- No storms: runner.attentionSent records sequence:event per run before publishing,
so a run that stays waiting is never republished on a re-request or restart, and
a stage notify policy that already covers the same event marks it first so the
run-level default is suppressed. The set is dropped when the run is terminal. A
delivery failure is logged, never fails the run, and is not retried.
- The payload names the run and pipeline and carries a click action
pipelines://run/<id> (validated: absolute URL, no userinfo, no fragment, no
credentials). PublishNtfy gained the optional click parameter; existing callers
pass "".
App:
- AndroidManifest/MainActivity declare the pipelines://run/<id> deep link
(singleTop); runIdFromDeepLink parses it and PipelinesApp opens that run once,
clearing the link. It opens a run; it starts nothing.
- ui/Attention.kt is the pure model (status -> Decision/Failed/Finished or
nothing, ordered by kind then newest-first) and ui/components/AttentionSurface.kt
is the floating indicator plus bottom sheet. It is derived from the same run
events the server already pushes, so an entry appears with no refresh and clears
when resolved; a finished run clears once looked at, and any entry can be
dismissed, persisted in SettingsRepository. The existing destinations, pill and
bottom strip are untouched and no competing navigation surface is added.
Checked before committing:
- Packaged authoring skill references: internal/author/skill/references/
notifications.md documents the automatic run-attention notifications and the
app surface. The only packaged "template" carrying notify is that reference; no
pipeline definition template exists to update and install.go's managed file list
is unchanged (notifications.md already existed).
- Repository guidelines: docs/ntfy.md gains an "Automatic run-attention
notifications" section and no longer claims "No block means no notifications";
the top-level README phone-notifications note now says a waiting/failed run
notifies automatically; android/README.md gains the dated feature section and
supersedes the previous section's note that the attention surface was absent.
- docker.md, install docs and the CLI/TUI docs describe mechanisms this change
does not alter; left as they are.
Tests: go test ./internal/server/... ./internal/stage/... ./internal/author/...
all ok; notify_test.go covers the event mapping and the real HTTP delivery
(destination, topic, click, no duplicate). Android :app:testDebugUnitTest ->
BUILD SUCCESSFUL, 58 suites / 531 tests / 0 failures (new AttentionTest, 13 cases,
covering ordering, appearing from a push, clearing on resolution, dismissal, the
empty case, persistence and the deep link). Device screenshots and the phone's own
notification are visual review's per the brief's device policy.
Scope: android/ plus the smallest server change; no new dependency, no run started
from the app, no installer or Docker change.
Treat a failure to reach the harness before any prompt or tool call is
admitted (connecting or creating the session) as a distinct retryable
reach failure, so a transport blip at the first stage no longer ends a
mutating run; keep post-admission readback/poll/agent failures retried
only on an explicit read-only stage and never replay admitted tool work.
Preserve the true cause when the context is already done (join ctx.Err
with the typed transport error) and report a sanitized unknown provider
error name/status instead of a blanket summary.
Make the isolated format-repair deadline proportional to the stage
timeout (floored, capped and clamped to remaining minus a margin) so a
slow-but-repairable result is recovered without ever extending the stage
deadline; the bounded ladder, value guard, deny-all formatter session,
cumulative usage accounting and no-tool-replay rule are unchanged.
Update docs/model-results.md, docs/pipeline-performance-review.md and the
packaged skill references; add stage, opencode and server fixtures.
chat_sessions used to list exactly the OpenCode sessions the pipelines own,
filtered through pipeline.SessionOwner (run_steps/run_session_owners), so an
operator-created chat was an accident of the run-ownership table and the
operator's working sessions were mixed with the internals of running pipelines.
Add a first-class operator session:
- internal/pipeline/operator_sessions.go: a dedicated operator_sessions table
(session_id PK, directory, title, tag, created_at, last_used_at) with
RecordOperatorSession (upsert preserving created_at), a one-statement
TouchOperatorSession (no-op for a non-operator id), OperatorSessions bounded to
128 rows most-recently-used first, and OperatorDirectories (bounded MRU,
deduped). run_steps/run_session_owners and SessionOwner/OwnedSessions are
untouched.
- internal/server/chat.go: chat_sessions returns operator sessions only - it
reads the registry once, intersects it with one broker session-list call, fills
directory/tag/created/updated from the registry, sets busy for a session held
by another connection, keeps MRU order, and stays fail-closed (empty list, never
the endpoint's whole list). A pipeline stage session and an untagged row can
never be listed. chat_create resolves the directory (empty/'~'/'~/...' ->
operator home; relative or missing rejected before any OpenCode call), calls
broker.Create(title, resolved) and records the row with the operator tag.
chat_directories returns home first then bounded (8) deduped recents. A
successful prompt touches the operator session without affecting pipeline
prompts.
- internal/opencode, internal/protocol, internal/server/websocket.go and
message.go: unscoped session list, Create posts {"title":...,
"location":{"directory":...}} to POST /api/session, additive protocol fields
(omitempty) and the ChatDirectories request/response, plus the OperatorRegistry
option defaulting to the first store. One OpenCode transport per server is
unchanged.
Docs and references: docs/cli.md gains a "Chat sessions" section documenting the
operator-only fail-closed list, directory defaulting/rejection, chat_directories
(home first, deduped, bounded 8) and why a dedicated request was chosen.
internal/author/skill/references/operations.md's stale "Brokered OpenCode chat"
text (chat_create unbrokered, run_steps filtering, "listed the owned scratch
session" evidence) is rewritten to the operator-session registry, the fail-closed
filter, directory defaulting and the corrected live-test wording. No new
reference file, so internal/author/install.go's inventory is unchanged.
Checked for staleness elsewhere: README.md, the docs/ topic pages and the
packaged server/TUI templates carry no chat-session ownership claims, so none
needed updating.
Verified: gofmt -l clean; go vet and go build ./... clean; go test -count=1
./internal/pipeline/... ./internal/protocol/... ./internal/opencode/...
./internal/server/... and the full go test -count=1 ./... pass.
Combines lane-go's typed omni search, launch-time model overrides, TUI
artifacts view, run-state websocket push, and the "reflect the change where
it belongs" committer rule with the Templates gallery, chat/omni-search and
other reviewed work already on main. The 12 files both sides touched keep
both intents composed rather than dropping either side: the omni search keeps
the typed search contract (CLI `pipelines search` and the TUI popup with type
badges and marked snippets) alongside the run-list quick filter; cli.md and
ui-ux-review.md document both the search rework and the artifacts view;
orchestration.md, harnesses.md and operations.md keep the model-override and
reflect-where-it-belongs guidance; and the protocol/CLI/server/TUI code keeps
the new search, run-event, override and artifacts paths alongside the existing
run, title and help surfaces.
Checks before committing:
- go build ./... -- pass
- go vet ./... -- pass
- go test -count=1 ./internal/... ./cmd/... -- all pass
Doc check before committing: nothing needs updating. The lane's own docs and
references (docs/cli.md, docs/ui-ux-review.md, and the orchestration,
harnesses and operations skill references) already carry the new features.
format.md and the packaged templates already describe the line-oriented result
format and the human decision stage from the earlier lane-go merge. The root
README already describes the omni search popup. Model overrides, the search
command, run events and the artifacts view are runtime/CLI/TUI surfaces that
cli.md documents; no template, README or topic-doc edit is required by this
merge.
The installed OpenCode 2.0.20 exposes the authenticated SSE event firehose at
/api/event and the /api HTTP surface, but Detect selects backend "v2" and Open
returned no owned process/URL, so chat stayed disabled and every chat_* request
was answered with "chat is not configured on this server".
Add opencode.OpenStream/OpenChat: for either backend they start the single owned
opencode serve process, discover the loopback URL and verify the backend health
contract (/api/info for v2, /global/health for v1). main and the live test share
OpenChat so the selection cannot drift. Chat now streams real incremental
session.text.delta/session.reasoning.delta frames via the existing
Connection.EventStream + NormalizeChatEvent, with no polling and no fake
streaming.
- Authenticate the owned endpoint with the per-backend user (opencode for v2,
pipelines for v1); the v2 server ignores OPENCODE_SERVER_USERNAME.
- Send the V2 prompt shape {"text":...}, which the v1 translation also accepts.
- Skip firehose event types the adapter does not surface before decoding their
data, so session.instructions.updated (delta is an object) no longer fails the
stream.
Unit tests cover the per-backend user, the prompt shape/auth, the captured real
V2 event shapes and the selection guards. An opt-in live test drives
chat_sessions/chat_messages/a completed chat_prompt/an interrupted chat_prompt
against the installed OpenCode. Update the packaged operations reference with the
realized capability, dated evidence and deployment state.
Checks before commit: the packaged authoring reference
internal/author/skill/references/operations.md is updated in this change; no new
reference file was added, so internal/author/install.go's managed inventory is
unchanged and the internal/author tests pass. No packaged template describes the
brokered chat surface, and neither README.md nor the topic docs under docs/ (nor
web/) document it, so none of them needed updating.
Bring the packaged starters and the pipelines skill up to the guidelines the
project now works to.
Templates:
- task and three-stage now end in a committer stage that stages the union of the
runner's captured input.server_changed_files and the producer's declared
changed_files, runs the repository's checks, commits locally and never pushes;
their producers no longer commit.
- Every model/agent prompt states the line-oriented result contract (outcome on the
first non-empty line, "# output" JSON map, other "# <heading>" sections kept
verbatim as evidence) and an "# evidence" section for exact commands and results,
with proportionate verification and no fabricated checks.
- Stages that can write save their reports, logs and screenshots into the runtime
artifacts directory.
- Remove the summary starter: a tool-free single-prompt text transform with no
repository interaction, artifact or commit is a format demonstration, not a
reusable workflow. No starter invents a pretend approval stage; a decision point
is a type=human stage.
Skill:
- references/format.md gains a Result format section (line form, legacy JSON
precedence, result_parse, raw-output.md, artifacts) and a Committer stages
section with the pre-commit checklist; the role-split and envelope notes mention
the line form and the committer.
- orchestration.md points at the checklist and records that a genuine gate is a
human stage; operations.md and the SKILL.md index drop the removed starter and
name the result format and committer. No new reference, so the managed inventory
is unchanged.
Docs/tests: examples/README, README, docs/cli, docs/first-version-testing, the CLI
help example and the smoke script track the four starters; catalog, tool-role,
server, tui and pipeline tests are updated, with a new test pinning the committer
ending and server-captured change set.
Commands run: go build ./...; go vet ./...; go test ./...; python3 -m unittest
discover -s scripts -p 'test_*.py'; python3 scripts/smoke-first-install.py. All
passed. Templates inspected through an isolated CLI server (pipelines templates;
pipelines template KEY --source). No server, service or device was restarted, and
nothing was pushed.
The packaged authoring skill's orchestration reference now tells an author
to carry the check from the brief into the pipeline: before committing, the
committer verifies the change is reflected in the packaged templates, the
repository's guidelines (README and topic docs) and the packaged skill
references, updates anything stale, and states in the commit body what was
checked or that nothing needed updating.
- brief-writing section: add a "Where it belongs" item to the brief and
renumber the hard-constraints and publishing items.
- hard-constraints block: add a "Reflect the change where it belongs" bullet.
- Pipeline choice: extend the committer bullet with the same pre-commit
check, keeping the existing rule that the committer stages only the run's
changed files.
Checked before committing: the change is prose in the packaged skill
reference, so no template, README or topic-doc edit was needed; the eight
pipelines carrying a committer stage were republished with the same rule in
their committer prompt (not tracked in this repository); no other reference
is stale.
The Artifacts view re-clamped its scroll offset to the selected row on every
render, so the tail of the list - a trailing visit that recorded none (running
or finished) and its capture summary - could never be brought on screen, and
the image marker crowded out artifact names at narrow widths.
- Decouple the scroll offset from the selection: add PgUp/PgDn (and ctrl+u/
ctrl+d) paging, and make G/end select the last artifact and scroll to the
tail so a trailing visit stays reachable.
- Share artifactContent/clampArtifactTop/revealArtifactSelection between
selection and render so both agree on the same page.
- Show the [img] marker only while the width still leaves room for a usable
name, keeping columns aligned either way.
- Advertise the scroll keys in the artifacts footer help.
- Test the omitted-files capture wording, trailing-visit reachability via
G/end and page scroll, and narrow-width no-overflow; render both the top and
the scrolled tail in the env-gated live capture test.
- Update the packaged operations reference for the A artifacts view.
The committer staged only the implementer's self-declared changed_files,
so an incomplete declaration silently left part of the approved change
uncommitted. The runner already records the run's real changed paths by
comparing the working tree around each step; publish that set to the
committer.
- runState gains an unbounded, deduplicated, first-seen-ordered captured
path accumulator (capturedPaths/capturedSeen), distinct from the bounded
changedPaths sample used only by the UI-surface skip reason.
- recordChanges folds each captured FileChange path into it in the existing
single pass, so no second walk and no new I/O.
- publishServerChangedFiles writes the set into the input memory value as
input.server_changed_files, called before memoryInput so the binding
always resolves (an empty list when nothing was captured) and only stages
that select it receive it. rebuildRunState rebuilds it from persisted
completed steps, so a resumed run hands the committer the same evidence.
- Tests prove the committer receives the captured set distinct from and
additive to the implementer's declared changed_files, that an omitted
path is still present, and that the field is an empty list when nothing
was captured.
- control.md and orchestration.md document the reserved field and instruct
the committer to select input.server_changed_files alongside the declared
changed_files and stage their union.
Commands run: go build ./...; go vet ./...; go test -count=1
./internal/server/... ./internal/stage/...; go test -count=1 ./internal/author/...;
go test -count=1 -run TestCommitterReceivesServerCapturedChangedPaths ./internal/server/.
All passed. go.mod/go.sum unchanged.
Add `pipelines run --model PROVIDER/MODEL` and repeatable
`--stage-model STAGE=PROVIDER/MODEL`, resolved with precedence
per-stage > whole-run > the model pinned in the pipeline version.
- A shared pure resolver (`pipeline.ParseModelOverride` /
`pipeline.EffectiveModels`) owns the precedence and validation: every value
must be a `provider/model` pair (`stage.ValidModel`), every stage key must
exist in the pinned graph, an unknown key fails naming the valid keys, and a
per-stage key that names a non-model stage is refused.
- The server validates against the pinned version before admission (no run row
is ever inserted for an invalid request), persists `model`/`stage_models` on
the run and `provider`/`model` on each step, applies the effective pair to a
clone of the stage definition so the stage and its repair/formatter path share
it, overlays the returned run graph with the effective pair (presentation
only), and reuses the recorded overrides on retry.
- The run record echoes the overrides it was launched with and each step records
the provider/model it actually ran with; the TUI Evidence pane renders that
effective pair. The pinned pipeline version and `show --version N` are never
modified.
- CLI help, `docs/cli.md` and the packaged authoring skill references document
the flags and the precedence. A TUI override picker remains a documented
follow-up; the CLI is the deliverable.
Tests: pure precedence/validation table tests, store round-trip and
request-key-mismatch plus migration, server admission/application/overlay/retry
with a controlled executor, CLI flag/help/malformed-value tests, and a TUI
evidence test.
Integrates nine commits of reviewed work from the lane-go branch after the
automated resolver failed on this merge five times (each attempt dying on the
malformed-result defect whose fix is contained in this same merge).
Integration decisions:
- Search: kept main's shipped design (Store.Search + SearchResponse.Results) and
dropped lane-go's parallel, older attempt (an unused Message.Matches field, an
absent runs.search, and two partial TUI plumbing blocks). lane-go itself did not
compile: protocol.SearchMatch was undefined and model.go referenced searchNav,
requestSearch and applyRunNav, which exist nowhere. Those partial TUI blocks are
removed; the queued TUI omni-search item will implement the surface properly on
main's design.
- Both sides' additions are kept where they are independent: run titles and
provenance, the human decision stage and its CLI/TUI surface, the derived stage
verdict, the stale-client guard, the line-oriented result format, and main's
search command.
- The search key's help label is now "/ omni search"; the two tests asserting the
old label are updated rather than the label reverted.
- docs/cli.md and the packaged operations reference carry both sides' text.
Verified: go build ./... , go vet ./... and the full go test ./... suite pass.
Add references/work.md: one durable ledger per project (queue, active, done,
failed), the start/pause/resume/status operations, per-item directives, a
restart-safe queue driver, sleep autonomy and a conciseness rule. Index it from
SKILL.md and add it to the installed inventory.
Exchange a build/protocol identity so an older CLI can no longer talk to a
newer native server silently:
- protocol carries ProtocolVersion and every server frame (including the
unsolicited welcome) is stamped with the server build and protocol, so a
client that never pings still learns the server identity.
- buildinfo.Compare classifies the relationship as InSync/Warn/Refuse and
names both builds; the client Watch reports one warning per connection and
returns a hard error when the server speaks a newer, uninterpretable
protocol so the caller never acts on the frame.
- client.ReceiveNotice decodes the raw frame so an unmodeled top-level field
produces a one-time "unrecognized field(s)" notice instead of being dropped
silently (the case that hid the artifacts field).
- CLI routes warnings to stderr, keeps machine JSON stdout clean and fails
non-zero on a refused mismatch; ping/connection print the client/server
relationship. The TUI surfaces the same notices through the shared log.
- scripts/setup-native.py now warns when an older `pipelines` earlier on PATH
shadows prefix/bin/pipelines, naming both paths and versions, without
installing or modifying anything else.
- Docs, the packaged debugging skill reference and README no longer imply the
installed CLI always matches the server.
Tests cover the comparison table, unknown-field detection and caching, the
server identity stamping on pong and welcome, the CLI/TUI reporting paths and
the installer PATH shadowing warning.
A new registered stage type `human` (internal/stage/human.go) has no model,
agent, tools or configs: its required `prompt` block carries the operator
instructions and its non-error edges are the selectable options. It parses,
validates, renders, exports and imports like every other type (no parser
special-case, no config keys).
At runtime the runner parks the run instead of executing: the step and the run
persist `waiting` with self-contained {prompt, options} evidence, and the run
returns without starting later stages, capturing file diffs, scanning artifacts,
holding a goroutine/timer or applying a stage timeout. The wait is durable: a
restart leaves the waiting rows untouched (RecoverRuns excludes them) and a
rebuilt runtime over the same database still accepts the decision.
`pipelines decide <run-id> [stage] --option KEY [--feedback TEXT] [--json]` and
the TUI `c` picker resolve the decision from the pinned graph in one store
transaction, record the choice and feedback as the step's evidence and resume
along edges[KEY]. An unknown option is refused with the valid list, a stage that
is not waiting is refused, and a repeated decision for an already-decided visit
does not double-apply. Cancel and retry keep working while waiting and after a
decision; a human stage is never killed by a stage timeout or max_visits, and a
`notify` policy can fire an `on=waiting` event.
Docs: the authoring skill references (format.md, operations.md), docs/cli.md and
README describe the stage, the command and the TUI picker. Tests cover stage
validation, the parsing round-trip, the store transitions, a websocket wait/decide
integration, cancel/retry, waiting notifications and a persistence-layer rebuild
(second Open + newRunner over one database, no process restart).
One query now finds anything in the system over the wire and from the CLI.
- protocol: add Search/SearchResults message types and a SearchRequest
(query, optional kind filter, optional limit) / SearchResponse (typed,
ranked matches plus truncation and scan-bound counters) pair, with typed
payloads for pipeline, version, run, step/visit, field text and artifact.
- store: add Store.Search, a read-only query over pipelines,
pipeline_versions, pipeline_runs, run_steps (+step_control) and
step_artifacts. Ranks exact name (3) > name substring (2) > body text (1),
breaks ties by recency then a stable kind/target order, and reports the
matched field. Bounds work with a 50-default/200-max result cap, a 2000-row
per-kind scan limit and a 2000-step / 1 MiB field-text scan budget.
- terms: a bare status word or lone date expression (today, yesterday, this
week, YYYY-MM-DD) selects runs; anything else stays free text.
- server: dispatch protocol.Search to stores[0].Search; read-only and
unconfirmed.
- cli: add `pipelines search QUERY [--type KIND] [--limit N] [--json]`, help
text and a BROWSE line.
- store: add OpenReadOnly so the live-database check never runs DDL.
- docs: document the grammar, kinds, ranking and bounds in docs/cli.md,
scope the TUI omnisearch to pipelines in docs/ui-ux-review.md, and add a
search bullet to the packaged operations reference.
- tests: protocol round trip, store/ranking/term-parsing/limit/scan-bound
unit tests, a server end-to-end test, and an env-guarded live check.
No new dependencies; go.mod/go.sum unchanged.
Move a completed stage visit's derived verdict out of the Progress summary
and into the top of the selected stage's Result section, so the chosen
option, its concise reason and the referenced artifacts are visible where
the stage is inspected.
- Add stageVerdictLines, which renders one line "Verdict: <option> · <reason>"
(still stating the option when the reason is absent) plus a following
"Referenced artifacts:" line only when artifacts exist, and returns nil
for a nil verdict so nothing extra is shown.
- Prepend those lines in runSectionLines after the stage state and before
the stage output and evidence, without adding a section or changing the
panel layout.
- Update the TUI test to render through the run-detail view, asserting the
option, reason and artifact names appear, and that a nil verdict adds
nothing; document the TUI placement in the debugging skill reference.
Omni-search work already present in the working tree (including the
runDuration→protocol.Duration refactor in run_detail.go) is unrelated to
this change and is left uncommitted.
Each completed step now exposes a verdict: the chosen option (outcome -> target),
a bounded one-line reason with model/fallback/absent provenance, and the recorded
artifacts the reason refers to. The store persists only the model-generated line
in a new step_verdicts table and derives the fallback at read time (no backfill);
the server generates the line once at completion, best-effort; the TUI renders it
in the stage progress summary. The packaged debugging reference documents it.
Add a bounded, never-empty title to every run. `pipelines run --title "..."`
sets it explicitly and skips model generation; otherwise the server generates
it from the prompt through the shared summariser and falls back to the prompt's
first line, recording title_source (model/fallback/explicit) and title_model.
- protocol/store: RunSummary.Title plus Label(), a set_run_title message, title
columns with an idempotent migration, CreateRunTitled, and SetRunTitle renames.
- summarise: Title/GenerateTitle/DeriveTitle/ExplicitTitle bounded to TitleCap.
- runner: generate the title outside the admission lock and only for a request
that starts a run; a replay reuses the stored title and an explicit title
skips the model entirely.
- CLI: `--title` on run and a `title <run-id> "<title>"` rename command; human
`runs` and `inspect` show the title beside the pipeline name and version.
- TUI: the run list row, run detail header, Home run cards and last-run line,
and the pipeline Results line show the title via Label().
Tests cover storage/migration/fallback, explicit-title and rename, the
end-to-end CLI flow, and title rendering; CLI and operations docs updated.
A model stage whose final result was malformed failed the whole run even when
the work and evidence had survived, because the only recovery was an isolated
formatter call on a fixed 30-second deadline that could time out on its own
(run 55, ui_reviewer of pipeline 9 v6).
Add an ordered, deterministic local repair pass (internal/stage/result_repair.go)
that runs before any model call: (a) append missing closing delimiters via the
existing closeJSONContainers; (b) remove one unambiguous trailing comma before a
closing } or ]; (c) insert exactly the single comma/colon the JSON parser reported
at its byte offset. Every candidate must pass losslessSyntaxRepair(original,
candidate) and decodeModelResult before it is accepted, so no field value is
invented, removed, reordered or rewritten. Ambiguous payloads (an unescaped quote
or colon inside a string value, or a stray delimiter with no single completion)
return not-ok and fall through to the bounded model re-emit rather than guessing.
stageResultText still strips surrounding prose and a single code fence first;
model_prompt.go now also strips prose around a single balanced-but-malformed
object span so the deterministic pass sees just the object.
Replace the fixed 30-second formatter deadline with repairDeadline(payloadBytes,
remaining), which grows monotonically with payload size, has a floor, and is
clamped to the attempt's remaining budget minus a safety margin, so the stage
context remains the outer bound and a large result is not guaranteed to time out.
The one-request-per-attempt, 32 KiB bound and the deny-all, retained, new-session
property of RepairFormat are unchanged for OpenCodeClient and PiClient.
Record an explicit recovery status on every terminal path: repaired, partial or
unrecoverable, plus a byte offset. A candidate that is valid JSON but still fails
envelope validation is marked partial and never returned as a completed result;
the unrecoverable path retains the bounded original payload, the precise parse
error (including its byte offset) and surfaces the status in the FINAL RESULT
VALIDATION diagnostic.
safeToReplay is intentionally unchanged. The deterministic local repair and the
fresh deny-all, no-tools, retained formatter session are independent of the
stage's tool set and never re-prompt or replay the admitted agent session, so a
shell-enabled review stage gains malformed-result recovery without any replay of
admitted work. Risk: no change to mutation safety; the only added cost is one
bounded formatter call, already budgeted. Replayability is not extended for shell
stages.
Docs and packaged skill references (docs/model-results.md,
internal/author/skill/references/{format,debugging,verification}.md) describe the
three-tier recovery, the status words and the derived deadline; the safeToReplay
decision and its risk are recorded there alongside the commit.
Tests add mocked, offline coverage for the deterministic and partial paths, a
repairDeadline monotonicity/floor/clamp table, and a shell-stage regression
proving a recoverable malformed result does not fail the run or replay tool work.
chat_sessions now brokers GET /api/session but filters the global list to
the sessions this server recorded in run_steps, deduped by session id, and
fails closed to an empty list when no ownership source is configured. A
session another client holds (in-server claim, released on stream end or
disconnect) or an active run is marked busy; a second client's chat_prompt
for a held session is rejected before any OpenCode call. chat_create stays
unbrokered.
Add the server-side half of the brokered chat path (route B): the
opencode.Chat adapter forwards chat_prompt to POST
/api/session/{id}/prompt through the owned Connection.Call translation
and derives chat_event frames from the authenticated /api/event SSE
firehose (text and reasoning deltas, tool parts with their status
transitions, step and usage). Every pushed frame carries a stable
session and message id, with a last-message-id fallback so step/tool
events that omit the assistant message id stay attributable.
The /connect handler keeps per-session retained state (chatHub) capped at
1 MiB with least-recently-updated eviction at 128 turns, so a
reconnecting client re-syncs through chat_messages instead of replaying
deltas the OpenCode per-session log never resends. Prompt dispatch
subscribes before posting and buffers pre-ack frames to preserve order,
and cancels the stream on disconnect so no goroutine leaks. chat_sessions
and chat_create are declared but not yet brokered: they answer with an
explicit protocol error rather than the generic Pong.
Reuse the existing transport, validation and normalization; no second
OpenCode transport or credential path is introduced.
Tests: protocol golden round-trip, adapter normalization + v1/v2 Call
translation, and handler ack/stream/validation/resume/disconnect, plus a
real-adapter no-tool evidence harness.
Rewrite SKILL.md to state the stance (reconnoitre, then route all work
through Pipelines runs; manage and debug, never implement), keep the
existing authoring/operations index, and route to four new references.
Add references/orchestration.md (stance, work items, brief template,
hard constraints, pipeline choice, running and watching),
references/lanes.md (worktrees, devices, sync/merge, restart-replay
cursor fix), references/debugging.md (eight symptom/confirm/do
playbooks) and references/delivery.md (artifacts, LAN URLs, WIP naming,
install boundaries), each with a clearly labelled project example where
Android/device specifics apply.
Append orchestration verification habits to references/verification.md
and add the four new files to install.go's managed inventory.
Native installs left Docker-era harness profiles in harnesses.json, so every
authoring launch failed trying to reach a Docker service, and a retained draft
bound to the old Docker server was silently projected onto the configured one.
- harness: add Profile.Mode and VerifyDockerLink as the single mode/verification
helpers. `harness setup --host` configures or converts a host profile and
clears docker_directory; reconfiguring a Docker-linked profile re-verifies the
recorded directory still owns compose.yaml and fails naming it instead of
saving a link that cannot launch. Setup's summary and `harness list` (via the
new List/ListView shape) report the resulting mode and, for Docker, the
directory, without changing the stored Config schema.
- setup-native.py: after build/install-skill it writes host (no docker_directory)
profiles for the detected harnesses, defaulting to the detected OpenCode profile
rather than a stale pi default, taking models from --model or the existing valid
profile and never inventing one; it clears a stale Docker link and preserves
unrelated keys/profiles. Detection-only: no harness install, upgrade or download.
- draft: add Draft.Resolve, the shared mismatch helper, which projects a draft
onto the configured server and returns a dual-URL notice. CLI validate/submit
and the TUI draft browser use it so a foreign draft is never silently
retargeted; the TUI pre-wraps the notice to the panel width.
- Tests cover the Docker->host and host->host transitions, the missing-directory
failure, installer write/migration (Python), and the CLI/TUI notice, using the
real config file path in a temp dir.
- Update help text, README, docs and the packaged authoring skill reference to
describe host as the default and the Docker-linked case.
Fix four stage failure-handling and observability defects (A2, A3, A6, A9).
- A2: add safeToReplay(def) as the single definition of retryable post-admission
work - only an explicit tools block that selects neither edit nor shell. The
agent/pi WorkStarted guards now break only when the stage is not safe to
replay, so read-only stages genuinely retry a post-admission timeout while
mutating and legacy-bundle stages still never replay admitted work.
- A6: re-emit a malformed final result on an explicit read-only stage up to the
retry budget, preserving format_repairs evidence; mutating stages keep the
immediate return and their session.
- A3: timed-out stages retain usable evidence. waitForSessionIdle records one
bounded sanitized OperationEvidence on deadline; the agent deferred cleanup
does a bounded best-effort latest-assistant readback and settles knowable V2
usage; RetainSession adds a retained_session pointer to failed step
transcripts so the session is visible beneath the BoundTranscript threshold.
- A9: a stage outcome routed to #failure now yields a non-empty run diagnostic
("stage <name> returned outcome <outcome>") while preferring an existing step
diagnostic.
Adds focused stage and server tests for retry safety, malformed-result
re-emission, timeout evidence retention and run-level diagnostics; updates the
packaged skill references and docs/cli.md, docs/model-results.md.