feat(opencode): task signals — structured returns, terse completion, sparse context, wake-on-message

[iceteaSA]

Issue for this PR

Related to #19215 (agent-team coordination) — the wake-on-message piece. The other three features are new task-tool capabilities without a tracking issue.

Stacked PR. This sits on top of #32693 (session-to-session), which sits on #32517 → #32425 → #32192. Against dev the diff shows 52 commits / 122 files, but only the top 13 commits (36 files, +1798/-49) are new here — everything below is the #32693 stack. Review against #32693, or wait for the base stack to land. Of the four features, only wake-on-message depends on the base stack (it uses the Messaging service from #32192/#32517); task_return, terse completion, and sparse context are self-contained and would apply on dev directly.

Type of change

• New feature

What does this PR do?

Four controls over what flows in and out of a Task dispatch. All default to today's behavior.

• task_return tool — structured child results. A subagent calls task_return({ result: {...} }) to set a free-form JSON object (≤4KB) on its own session row; it lands in the completion frame and on a new optional result field of the task.completed event. This is the output symmetry to the metadata input param — reviewers/coordinators can read a machine result instead of regex-parsing a "VERDICT:" line out of prose. Oversize is a model-visible tool error, not a silent truncation.
• Terse completion frames — completion: "full" | "terse". A background subagent's full <task_result> body is injected verbatim into the parent, which rewrites the parent's cache tail every completion — on orchestration-heavy runs that's dozens of 2–6K-token cache busts. Terse mode replaces the frame body with a digest: the structured task_return result, the last 500 chars of the final message (verdict lines live at the tail), and a pointer to the child session for the full output. The parent can always resume the child to read the rest.
• Sparse dispatch context — context: "full" | "sparse". A general-style dispatch pays ~35K first-turn input tokens; ~12–14K of it is instructional payload a scoped child never uses. Sparse keeps the project AGENTS.md chain, the agent prompt, tool schemas, and env; it drops global instruction files, the skills block, and the MCP-instructions block — and genuinely skips assembling them, not just discards the output. Plugin system-prompt transforms still run (they're post-assembly), so plugins keep their say.
• Wake-on-message — wake_on_message: true. Today a sibling/coordinator message to an idle child sits undelivered until something else wakes it. With this flag, an idle child is woken to drain its inbox when a message lands. Opt-in per dispatch (an always-on wake reintroduces interrupt storms), budgeted at 5 wakes per run (refreshed when the parent resumes the child), and never wakes a child whose ancestor session chain is gone.

Terse and sparse both read from a new task config family (task.completion / task.context) with precedence global config → project config → agent frontmatter → dispatch param, so an orchestration-heavy setup sets the default once and overrides per dispatch.

How did you verify your code works?

Executed as six reviewed tasks (implement → independent cross-family review → fix), red-first throughout.

• Full suite green: bun test across test/tool/ test/session/ test/messaging/ test/s2s/ — 787 pass / 0 fail; typecheck clean on opencode, core, schema, tui. Two migrations (result, context_mode), each column mapped through all six parallel session-row sites (V1 fromRow/toRow, projector, V1 SessionInfo, V2 fromRow, V2 Info) — a missed site is silent data loss, which is exactly what the T1 review caught on the V2 read path before it shipped.
• Wake-on-message was the risk, and review earned its keep. The first implementation forked the wake via a bare Effect.runFork, which runs against the default runtime and does not carry the per-request InstanceRef the Messaging/SessionStatus services resolve through — so prompt.loop would die on a missing-InstanceRef defect in production while every spy-handler unit test passed green. A reviewer proved it with a live probe (child-ref= MISSING) and rejected it. The fix captures the instance context with attach(...).pipe(Effect.forkIn(scope)) (the same pattern the s2s poller uses) and adds an integration test that stands up the real SessionPrompt layer and drives a real prompt.loop drain — it goes red if either the registration or the context-carrying fork is reverted.
• Mutation-checked seams: the terse-frame branch, the sparse skills/mcp-drop branch, and the wake idle-predicate each fail their test when reverted.

Honest limit, same as #34947's fallback caveat: the token-economy wins (terse cache-bust reduction, sparse first-turn token drop) and the end-to-end wake round-trip are covered by unit/integration tests and the instance-context probe, but the before/after token measurements are pending a live build — the tool-test harness can't measure real cache-tail rewrites.

Screenshots / recordings

Not a UI change.

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR (the base-stack commits in the diff are the declared feat(opencode): session-to-session messaging — communicate between two running sessions #32693 dependency — see the note at the top)