feat: harden autonomy lifecycle, OOM bounds, and provider-boundary finalization

This PR consolidates a coordinated batch of fixes around autonomy run/flow lifecycle, scheduled task deduplication, provider-boundary state finalization, and matching memory-bound treatments for adjacent long-running subsystems (REPL fullscreen scrollback, skill-search/skill-learning runtime activation). All changes were developed and reviewed together because they touched the same lifecycle invariants and were uncovered by the same long-running session reproductions.

## Lifecycle correctness

- Queued autonomy prompts are not injected unless the persisted run was successfully claimed; queued run claiming is now terminal-safe so a once-consumed/cancelled/failed run can not slip back into `queued`.
- Autonomy run/flow finalization happens on completion, provider error, generator close, and cancellation — not just the happy path. New `src/__tests__/queryAutonomyProviderBoundary.test.ts` covers these provider-boundary transitions.
- `requestManagedAutonomyFlowCancel` and `resumeManagedAutonomyFlowPrompt` carry `rootDir` and `currentDir` explicitly across detached async boundaries (proactive-tick, cron, daemon restart) instead of inferring from process state.
- Active runs/flows are protected from janitor pruning so a running step can not be garbage-collected mid-flight (`src/utils/autonomyAuthority.ts`).
- Heartbeat parser now ignores fenced code blocks; the two-phase commit window for autonomy state transitions is documented in `docs/internals/autonomy-jira.md`.

## Ownership and dedup

- `src/utils/autonomyRuns.ts`: ownership stamping (run id + rootDir carried end-to-end), source-based dedup against active runs.
- `src/hooks/useScheduledTasks.ts`: scheduled ticks deduplicate against runs already active on the same source label.
- `src/utils/processUserInput/processSlashCommand.tsx`: forked slash commands now thread the autonomy `runId` so completion finalizers can find the originating run for deferred completion.
- New `src/utils/autonomyQueueLifecycle.ts` and tests collect the queue-side lifecycle invariants in one place.

## Memory bounds (related, same review pass)

- `src/screens/REPL.tsx`: caps fullscreen scrollback after the compact boundary and updates trailing progress rows in place. Long-running fullscreen sessions could otherwise retain thousands of post-compaction messages and duplicate progress rows, keeping Ink trees alive long after their useful context had moved on.
- `src/services/skillSearch/*` and `src/services/skillLearning/*`: runtime activation is strictly opt-in via existing env toggles; session caches are capped so long-running processes can not grow them forever. Build presence is preserved so operators can still discover and opt into the slash commands.

## CI / test contract

- `tests/integration/dependency-overrides.test.ts`: smoke test no longer drives Mermaid's browser renderer; it validates the package-resolution contract directly so CI does not regress on unrelated browser timing.
- New `tests/integration/autonomy-lifecycle-user-flow.test.ts`: end-to-end CLI subprocess flow exercising `status --deep`, `flows`, `flow <id>`, `flow resume`, `flow cancel` against persisted state.
- `src/entrypoints/cli.tsx`: `claude autonomy …` routes through an entrypoint fast path that reuses the slash-command formatter without booting the full interactive CLI. Stdout is flushed before forced exit so coverage subprocesses do not terminate with empty stdout.
- `packages/builtin-tools/src/tools/RemoteTriggerTool/__tests__/RemoteTriggerTool.test.ts`: stabilized to prevent audit flake under coverage.

## Tests added

- `src/__tests__/queryAutonomyProviderBoundary.test.ts`
- `src/hooks/__tests__/useScheduledTasks.test.ts`
- `src/utils/__tests__/autonomyAuthority.test.ts`
- `src/utils/__tests__/autonomyFlows.test.ts` (extended)
- `src/utils/__tests__/autonomyPersistence.test.ts` (extended)
- `src/utils/__tests__/autonomyQueueLifecycle.test.ts`
- `src/utils/__tests__/autonomyRuns.test.ts` (extended)
- `src/utils/processUserInput/__tests__/processSlashCommand.test.ts`
- `tests/integration/autonomy-lifecycle-user-flow.test.ts`

## Docs

- `docs/agent/sur-loop-scheduled-oom.md`: System Understanding Report covering the scheduled/loop OOM problem, the call graphs investigated, and the lifecycle invariants this PR establishes.
- `docs/agent/sur-skill-overflow-bugs.md`: SUR for the related skill-overflow context.
- `docs/internals/autonomy-jira.md`: documents the two-phase commit window and ownership stamping invariants.
- `docs/memory-leak-audit.md`: audit notes covering the REPL/scrollback and skill-search bounds.

## Invariants this PR establishes

1. Queued autonomy prompts are not injected unless the persisted run was successfully claimed.
2. Terminal run/flow states are terminal — completion, failure, and cancellation all finalize state regardless of which provider/error path triggered them.
3. Autonomy run/flow `rootDir` is carried explicitly across detached async boundaries instead of inferred from a shared singleton.
4. State-only CLI subcommands (`autonomy status|runs|flows|flow …`) bypass full interactive bootstrap so they do not hold unrelated handles open.
5. REPL fullscreen scrollback and skill-search/skill-learning session caches are explicitly bounded.

## Validation

```bash
bun run typecheck
CI=true GITHUB_ACTIONS=true bun test            # 3996 pass / 0 fail across 305 files
bun test src/__tests__/queryAutonomyProviderBoundary.test.ts \
         src/hooks/__tests__/useScheduledTasks.test.ts \
         src/utils/__tests__/autonomy{Runs,Flows,Authority,QueueLifecycle,Persistence}.test.ts \
         src/utils/processUserInput/__tests__/processSlashCommand.test.ts \
         tests/integration/autonomy-lifecycle-user-flow.test.ts
```

## Origin

This PR is the consolidated, upstream-targeted version of two fork-side review PRs (fix/loop-scheduled-autonomy-oom and fix/autonomy-lifecycle). The fork-side review history is preserved at https://github.com/amDosion/claude-code-bast/pull/7 . The fork's own internal `chore: keep fork current with upstream` sync commits and the `docs: update contributors` automation are intentionally not included in this PR.

The autonomy CLI handler `rootDir` threading that the fork added (78f64d8a, 98d04ddb) is intentionally omitted here because upstream `a2cfaf91` (fix: 修复 RemoteTriggerTool 和 autonomy 测试的全量运行失败) already performed the equivalent change with an additional `currentDir` option. Keeping the upstream version avoids regressing that improvement.

src/services/skillLearning/featureCheck.ts

@@ -1,12 +1,36 @@
 import { feature } from 'bun:bundle'
 
-export function isSkillLearningEnabled(): boolean {
-  if (process.env.SKILL_LEARNING_ENABLED === '0') return false
-  if (process.env.SKILL_LEARNING_ENABLED === '1') return true
-  if (process.env.FEATURE_SKILL_LEARNING === '0') return false
-  if (process.env.FEATURE_SKILL_LEARNING === '1') return true
-  if (feature('SKILL_LEARNING')) {
-    return true
-  }
+/**
+ * Build-time presence check: is the `/skill-learning` slash command
+ * compiled into this build? Used by the command registry's `isEnabled` so
+ * the command appears in the menu whenever it is buildable. Operators
+ * activate the subsystem itself via `/skill-learning start`, which flips
+ * `SKILL_LEARNING_ENABLED=1` and turns the runtime observers on (see
+ * `isSkillLearningEnabled`).
+ */
+export function isSkillLearningCompiledIn(): boolean {
+  if (feature('SKILL_LEARNING')) return true
+  return false
+}
+
+/**
+ * Runtime activation check: is the skill-learning subsystem actively
+ * running (toolEvent, runtime, session observers attached, persisting
+ * observations to disk)? Off by default — the operator must run
+ * `/skill-learning start` (which sets `SKILL_LEARNING_ENABLED=1`).
+ *
+ * Legacy `FEATURE_SKILL_LEARNING=1` is also accepted for backward
+ * compatibility with operators who set it before the slash-command UX
+ * landed.
+ *
+ * Build-flag gating is intentionally NOT performed here: the command
+ * registry already gates command compilation on the build flag, and this
+ * function is only reached from code paths that the build flag has
+ * already let through. Decoupling keeps the test surface clean (tests
+ * exercise the env-var contract without needing to mock `bun:bundle`).
+ */
+export function isSkillLearningEnabled(): boolean {
+  if (process.env.SKILL_LEARNING_ENABLED === '1') return true
+  if (process.env.FEATURE_SKILL_LEARNING === '1') return true
   return false
 }

src/services/skillLearning/projectContext.ts

@@ -45,15 +45,44 @@ export function getProjectContextPath(projectId: string): string {
 // in the tool.call hot path (one wrapper invocation per tool) that cost would
 // accumulate into the hundreds-of-ms range per session. Cache keyed by the
 // exact cwd string so different worktrees still get independent entries.
+//
+// Bounded with LRU eviction: long-lived processes that traverse many
+// worktrees (e.g. multi-repo build orchestrators) would otherwise grow the
+// cache without limit. Each entry holds a SkillLearningProjectContext
+// (instinct + skill lists), so the cap ensures bounded memory regardless
+// of cwd diversity. `defines.ts` originally flagged this as
+// "无淘汰机制（非 GB 级主因）" — this fix closes that gap.
+const PROJECT_CONTEXT_CACHE_MAX = 32
+const PROJECT_CONTEXT_CACHE_TRIM_TO = 24
 const contextCache = new Map<string, SkillLearningProjectContext>()
 const PERSIST_INTERVAL_MS = 5 * 60 * 1000
 let lastPersistAt = 0
 
+function setProjectContextCache(
+  cwd: string,
+  ctx: SkillLearningProjectContext,
+): void {
+  if (contextCache.has(cwd)) contextCache.delete(cwd)
+  contextCache.set(cwd, ctx)
+  if (contextCache.size > PROJECT_CONTEXT_CACHE_MAX) {
+    const toDrop = contextCache.size - PROJECT_CONTEXT_CACHE_TRIM_TO
+    const iter = contextCache.keys()
+    for (let i = 0; i < toDrop; i++) {
+      const next = iter.next()
+      if (next.done) break
+      contextCache.delete(next.value)
+    }
+  }
+}
+
 export function resolveProjectContext(
   cwd = process.cwd(),
 ): SkillLearningProjectContext {
   const cached = contextCache.get(cwd)
   if (cached) {
+    // Refresh insertion order so frequently-accessed cwds survive eviction.
+    contextCache.delete(cwd)
+    contextCache.set(cwd, cached)
     // Still touch the registry so long-lived processes keep `lastSeenAt`
     // reasonably fresh, but throttle the write so it doesn't fire on every
     // tool call.
@@ -65,7 +94,7 @@ export function resolveProjectContext(
     return cached
   }
   const resolved = resolveContext(cwd)
-  contextCache.set(cwd, resolved)
+  setProjectContextCache(cwd, resolved)
   persistProjectContext(resolved)
   lastPersistAt = Date.now()
   return resolved

src/services/skillLearning/promotion.ts

@@ -23,8 +23,30 @@ export type PromotionOptions = {
   minConfidence?: number
 }
 
+/**
+ * Set bounded with FIFO eviction. # promotions per session is small in
+ * practice (single digits), but a long-lived sandbox/daemon could push
+ * this if it never restarts. The cap is defensive and the degraded
+ * behaviour — re-promote if we exceed N then forget the oldest — is
+ * benign because promotion is idempotent at the lifecycle layer.
+ */
+const SESSION_PROMOTED_IDS_MAX = 256
+const SESSION_PROMOTED_IDS_TRIM_TO = 192
 const sessionPromotedIds = new Set<string>()
 
+function recordSessionPromoted(id: string): void {
+  sessionPromotedIds.add(id)
+  if (sessionPromotedIds.size > SESSION_PROMOTED_IDS_MAX) {
+    const toDrop = sessionPromotedIds.size - SESSION_PROMOTED_IDS_TRIM_TO
+    const iter = sessionPromotedIds.values()
+    for (let i = 0; i < toDrop; i++) {
+      const next = iter.next()
+      if (next.done) break
+      sessionPromotedIds.delete(next.value)
+    }
+  }
+}
+
 export function resetPromotionBookkeeping(): void {
   sessionPromotedIds.clear()
 }
@@ -103,7 +125,7 @@ export async function checkPromotion(
     }
     await saveInstinct(globalInstinct, globalOptions)
 
-    sessionPromotedIds.add(candidate.instinctId)
+    recordSessionPromoted(candidate.instinctId)
     promoted.push(candidate)
   }