@ai-dossier/sched 0.42.1 → 0.45.0

package/README.md

@@ -665,7 +665,35 @@ below. 1.9.0 and earlier states migrate on load, backfilling absent OR explicit
 those same defaults: nothing before this field existed was ever weighted differently, so
 the backfill is exact, not a guess.
 
-## Unit priority (#565)
+## Per-member review level (#771)
+
+A slot member carries `review: 'light' | 'full'` on its `QueueEntry` (default `light`;
+pre-#771 `state.json` entries backfill to `light` on load — exact, since every earlier slot
+member got the light, relevance-scoped review). `review=full` is how a risk-floor issue
+(billing/security/auth/deploy/migration) joins a batch instead of being forced out to
+`mode=full` (#770, operator Option A):
+
+- **Enqueue** — set per manifest entry (`"review": "full"`) or with `sched enqueue --mode
+  slot --review full`. Rejected on a `mode=full` entry (a full cycle always reviews fully).
+  At most `MAX_FULL_REVIEW_MEMBERS` (2) `review=full` members per batch, counted over the
+  batch's whole membership after the call so an incremental join cannot slip a third in;
+  a violating call is rejected with an `EnqueueError` and writes nothing. The cap is
+  overridable per project via `config.json`'s `max_full_review_members` (a non-negative
+  integer), resolved by the CLI and passed to `enqueueEntries` as
+  `options.maxFullReviewMembers` — the core stays config-free.
+- **Dispatch** — `memberDispatchTier` floors a `review=full` member at `strong`, whatever
+  tier its manifest asked for; the recorded `QueueEntry.tier` stays as written. Only the
+  tier moves: the batch's dispatch profile still resolves the executor for that tier, so the
+  batch stays one agent family. The member's `spawned` journal event carries `review: "full"`.
+  The floor follows the member: its bounded fix attempt dispatches at `strong` (not
+  `FIX_ATTEMPT_TIER`), and an eviction/abandon requeue to full-cycle rewrites its `tier` to the
+  floored value (and `review` back to `light`, a slot-only concept) so the full cycle keeps it.
+- **Prompt** — `buildMemberPrompt` substitutes a `{review}` placeholder; a `review=full`
+  member whose template lacks one gets `FULL_REVIEW_MEMBER_DIRECTIVE` appended (full-cycle-
+  grade review before handover). A `light` member's prompt is byte-identical to before.
+- **Status** — `sched status`'s Queue table has a `review` column for slot members, and the
+  `tier` column shows the floor as `mid→strong` when it applies.
+
 
 `readiness.ts`'s `runnableUnits` ranks EVERY candidate — issues and batches together — by
 `priority` desc, then readiness age (`updated_at`) asc, then a numeric tiebreak (an
@@ -702,6 +730,17 @@ row governs assignment), so the Queue table renders `-` for it rather than a num
 looks load-bearing but is not. A batch dissolve (`recovery.ts`'s `dissolveBatch`) carries
 the parent batch's priority forward onto both split halves.
 
+## The batch gate — `gate.batch` (#777)
+
+`runValidate` runs the injected `runSuite(worktree, ctx)` once per `validating` pass;
+`ctx` is `{ batchId, baseRef: 'origin/<base_branch>' }` (`BatchSuiteContext`, also passed
+on attribution/fix re-runs). The CLI's runner (`cli/src/batch-suite-runner.ts`) prefers the
+repo's active `gate.batch` capability — the same CI-parity gate a normal PR pays — over
+`test.full`, exporting `ctx` as `DOSSIER_BATCH_ID` / `DOSSIER_BATCH_BASE`; outcomes map
+identically either way (`ok` / red-and-attributable / `suite-unreadable` block).
+`sched enqueue` refuses to form a batch whose only full gate is a `test.full` marked
+`timeout_prone: true`. Details: `docs/reference/capabilities.md`.
+
 ## Batch dispatch (#523)
 
 `batch-dispatch.ts`'s `runBatchTick` — called from `tick()` after the issue-level pass,
@@ -1188,6 +1227,24 @@ after-the-fact recovery, not a missing-data bug.
   can be manual (`sched pause`) or automatic (dispatch-health, #505/#629 above); `sched
   resume` clears the flag and both dispatch-health streaks.
 
+## Status health warnings (#776)
+
+`sched status` (and `warnings[]` in `--json`, built by `buildStatusWarnings`) flags state
+nobody came back to, each with its exact remedy:
+
+- **`long-pause`** — paused for more than 24 h (`STATUS_HEALTH_WARNING_AGE_MS`), measured
+  from `paused_at`, which `setPaused` stamps on the running → paused edge and clears on
+  resume. A pause recorded before `paused_at` existed is flagged as "unknown age".
+- **`stale-engine-lease`** — the engine lease holder's pid is dead while queue entries are
+  unfinished or slots live: nothing is ticking (`sched start`).
+- **`stuck-slot`** — a live slot (`assigned`/`running`/`recovering`) with no progress for
+  more than 24 h (`sched stop --issue N` / `--batch B` once the work is confirmed done or
+  abandoned).
+- **`stale-closed`** — each tick polls `issueClosed` for every `recovering` cycle slot; a
+  closed issue sets the entry's sticky `stale_closed_at`, journals `stale-closed` once, and
+  the recovery rail never respawns it (even after `sched resume`). Report slots are exempt:
+  their issue is closed at merge by design. Remedy: `sched stop --issue N`.
+
 ## Hard-block labels are re-read every tick (#544)
 
 #507's enqueue pre-screen resolves an issue's GitHub labels in the CLI and lands the

package/dist/batch-dispatch.d.ts

@@ -65,7 +65,7 @@ import { type GroundTruth } from './groundtruth';
 import { type Journal } from './journal';
 import type { SchedStore } from './persist';
 import type { ExecFn } from './project';
-import { type SuiteResult } from './recovery';
+import { type BatchSuiteContext, type SuiteResult } from './recovery';
 import { type FsExists } from './teardown';
 import type { BatchEntry, CapabilityGateResult, SchedConfig } from './types';
 export type { CapOutcome } from './types';
@@ -88,8 +88,13 @@ export interface BatchDispatchDeps {
      * timeout, e.g. `@ai-dossier/worktree-pool`'s `WARM_COMMAND_TIMEOUT_MS`).
      */
     warmExec?: ExecFn;
-    /** Runs the aggregate suite inside a batch worktree; batches never leave `validating` without one. */
-    runSuite: (worktree: string) => SuiteResult;
+    /**
+     * Runs the aggregate suite inside a batch worktree; batches never leave
+     * `validating` without one. `ctx` (#777) names the batch and its base ref so
+     * the runner can hand them to the repo's `gate.batch` capability; optional
+     * so runners written before it keep working unchanged.
+     */
+    runSuite: (worktree: string, ctx?: BatchSuiteContext) => SuiteResult;
     /**
      * Runs one `ai-dossier cap run <capabilityId>` in a batch worktree. Two call
      * sites, different degrade contracts:

package/dist/batch-dispatch.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"batch-dispatch.d.ts","sourceRoot":"","sources":["../src/batch-dispatch.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6DG;AAmBH,OAAO,EAYL,KAAK,gBAAgB,EAGrB,KAAK,SAAS,EACf,MAAM,YAAY,CAAC;AAEpB,OAAO,EACL,KAAK,WAAW,EAQjB,MAAM,eAAe,CAAC;AACvB,OAAO,EAAE,KAAK,OAAO,EAAa,MAAM,WAAW,CAAC;AACpD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAC5C,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AAExC,OAAO,EAUL,KAAK,WAAW,EACjB,MAAM,YAAY,CAAC;AAoBpB,OAAO,EAAE,KAAK,QAAQ,EAA2D,MAAM,YAAY,CAAC;AACpG,OAAO,KAAK,EAEV,UAAU,EAEV,oBAAoB,EAMpB,WAAW,EAGZ,MAAM,SAAS,CAAC;AAYjB,YAAY,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAE1C,8DAA8D;AAC9D,MAAM,WAAW,iBAAiB;IAChC,KAAK,EAAE,UAAU,CAAC;IAClB,OAAO,EAAE,OAAO,CAAC;IACjB,WAAW,EAAE,WAAW,CAAC;IACzB,SAAS,EAAE,SAAS,CAAC;IACrB,GAAG,EAAE,MAAM,IAAI,CAAC;IAChB,uFAAuF;IACvF,OAAO,EAAE,MAAM,CAAC;IAChB,0FAA0F;IAC1F,IAAI,EAAE,MAAM,CAAC;IACb;;;;;;OAMG;IACH,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,sGAAsG;IACtG,QAAQ,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,WAAW,CAAC;IAC5C;;;;;;;;;;;;;;;;;;;OAmBG;IACH,aAAa,CAAC,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,KAAK,oBAAoB,CAAC;IACjF,QAAQ,CAAC,EAAE,QAAQ,CAAC;IACpB;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,8FAA8F;AAC9F,MAAM,WAAW,eAAe;IAC9B,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,SAAS,EAAE,MAAM,EAAE,CAAC;IACpB,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,uFAAuF;IACvF,qBAAqB,EAAE,MAAM,EAAE,CAAC;IAChC,8FAA8F;IAC9F,OAAO,EAAE,MAAM,EAAE,CAAC;CACnB;AAED;;;;;;GAMG;AACH,MAAM,WAAW,aAAa;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAClC;AA0hBD;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,MAAM,CAE3F;AAswCD;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,wBAAgB,sBAAsB,CACpC,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,EAC1B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,UAAU,EACjB,WAAW,EAAE,MAAM,EACnB,OAAO,EAAE,aAAa,EACtB,GAAG,EAAE,IAAI,EACT,MAAM,EAAE,eAAe,GACtB,IAAI,CA+BN;AAiSD;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,wBAAgB,iBAAiB,CAC/B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,EAC1B,OAAO,EAAE,MAAM,EACf,GAAG,EAAE,IAAI,GACR;IACD,OAAO,EAAE,eAAe,GAAG,SAAS,GAAG,WAAW,GAAG,SAAS,CAAC;IAC/D,UAAU,EAAE,MAAM,CAAC;IACnB;;;;;OAKG;IACH,SAAS,CAAC,EAAE,cAAc,GAAG,qBAAqB,CAAC;IACnD,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB,CA2KA;AA+gCD;;;;;;;;;;;;;GAaG;AACH,wBAAgB,YAAY,CAC1B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,GACzB,eAAe,CA+PjB"}
+{"version":3,"file":"batch-dispatch.d.ts","sourceRoot":"","sources":["../src/batch-dispatch.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6DG;AAmBH,OAAO,EAaL,KAAK,gBAAgB,EAGrB,KAAK,SAAS,EACf,MAAM,YAAY,CAAC;AAEpB,OAAO,EACL,KAAK,WAAW,EAQjB,MAAM,eAAe,CAAC;AACvB,OAAO,EAAE,KAAK,OAAO,EAAa,MAAM,WAAW,CAAC;AACpD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AAC5C,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,WAAW,CAAC;AAExC,OAAO,EACL,KAAK,iBAAiB,EAUtB,KAAK,WAAW,EACjB,MAAM,YAAY,CAAC;AAoBpB,OAAO,EAAE,KAAK,QAAQ,EAA2D,MAAM,YAAY,CAAC;AACpG,OAAO,KAAK,EAEV,UAAU,EAEV,oBAAoB,EAOpB,WAAW,EAGZ,MAAM,SAAS,CAAC;AAYjB,YAAY,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AAE1C,8DAA8D;AAC9D,MAAM,WAAW,iBAAiB;IAChC,KAAK,EAAE,UAAU,CAAC;IAClB,OAAO,EAAE,OAAO,CAAC;IACjB,WAAW,EAAE,WAAW,CAAC;IACzB,SAAS,EAAE,SAAS,CAAC;IACrB,GAAG,EAAE,MAAM,IAAI,CAAC;IAChB,uFAAuF;IACvF,OAAO,EAAE,MAAM,CAAC;IAChB,0FAA0F;IAC1F,IAAI,EAAE,MAAM,CAAC;IACb;;;;;;OAMG;IACH,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB;;;;;OAKG;IACH,QAAQ,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,CAAC,EAAE,iBAAiB,KAAK,WAAW,CAAC;IACrE;;;;;;;;;;;;;;;;;;;OAmBG;IACH,aAAa,CAAC,EAAE,CAAC,QAAQ,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,KAAK,oBAAoB,CAAC;IACjF,QAAQ,CAAC,EAAE,QAAQ,CAAC;IACpB;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,8FAA8F;AAC9F,MAAM,WAAW,eAAe;IAC9B,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,SAAS,EAAE,MAAM,EAAE,CAAC;IACpB,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,uFAAuF;IACvF,qBAAqB,EAAE,MAAM,EAAE,CAAC;IAChC,8FAA8F;IAC9F,OAAO,EAAE,MAAM,EAAE,CAAC;CACnB;AAED;;;;;;GAMG;AACH,MAAM,WAAW,aAAa;IAC5B,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAClC;AAmiBD;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,MAAM,CAE3F;AA4wCD;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,wBAAgB,sBAAsB,CACpC,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,EAC1B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,UAAU,EACjB,WAAW,EAAE,MAAM,EACnB,OAAO,EAAE,aAAa,EACtB,GAAG,EAAE,IAAI,EACT,MAAM,EAAE,eAAe,GACtB,IAAI,CA+BN;AAiSD;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,wBAAgB,iBAAiB,CAC/B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,EAC1B,OAAO,EAAE,MAAM,EACf,GAAG,EAAE,IAAI,GACR;IACD,OAAO,EAAE,eAAe,GAAG,SAAS,GAAG,WAAW,GAAG,SAAS,CAAC;IAC/D,UAAU,EAAE,MAAM,CAAC;IACnB;;;;;OAKG;IACH,SAAS,CAAC,EAAE,cAAc,GAAG,qBAAqB,CAAC;IACnD,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB,CA2KA;AAkhCD;;;;;;;;;;;;;GAaG;AACH,wBAAgB,YAAY,CAC1B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,GACzB,eAAe,CA+PjB"}

package/dist/batch-dispatch.js

@@ -356,9 +356,10 @@ function applyBatchAndIssues(fresh, computed, batchId, issues, membersBeforeReco
  * into `tick()`'s own catch — a bare `tick-failed` with no unit id, repeating
  * every reconcile interval forever since nothing about the batch changed.
  */
-function safeSuite(deps, batchId, worktree) {
+function safeSuite(deps, batch, worktree) {
+    const batchId = batch.id;
     try {
-        return deps.runSuite(worktree);
+        return deps.runSuite(worktree, batchSuiteContext(batch));
     }
     catch (err) {
         const detail = `suite runner threw: ${err.message}`;
@@ -368,13 +369,19 @@ function safeSuite(deps, batchId, worktree) {
         return { ok: false, failing: [], readable: false, detail };
     }
 }
+/** What the aggregate-suite runner is told about the batch it gates (#777). */
+function batchSuiteContext(batch) {
+    return { batchId: batch.id, baseRef: `origin/${batch.base_branch}` };
+}
 function recoveryDeps(deps, config, batch, now) {
     return {
         exec: deps.exec,
         repoDir: batch.worktree ?? deps.repoDir,
         journal: deps.journal,
         postMilestone: (0, recovery_1.createExecMilestonePoster)(deps.exec, { repoDir: deps.repoDir }),
-        runSuite: batch.worktree !== null ? () => deps.runSuite(batch.worktree) : undefined,
+        runSuite: batch.worktree !== null
+            ? () => deps.runSuite(batch.worktree, batchSuiteContext(batch))
+            : undefined,
         dissolvePolicy: (0, types_1.resolveDissolvePolicy)(config.dissolve_policy),
         now: () => now,
     };
@@ -914,13 +921,17 @@ function spawnMember(deps, dispatch, state, slot, batchId, now, result, member)
         return releaseSlot(state, batchId, now);
     }
     const withStatus = advanceMemberToInWork(state, memberIssue, now);
-    const tier = (0, state_1.findEntry)(withStatus, memberIssue)?.tier ?? 'mid';
+    const memberEntry = (0, state_1.findEntry)(withStatus, memberIssue);
+    // #771: a `review=full` member dispatches at `strong` minimum (within this
+    // batch's profile — `dispatch` is already the batch-profiled resolution).
+    const tier = memberEntry ? (0, dispatch_1.memberDispatchTier)(memberEntry) : 'mid';
+    const review = memberEntry?.review ?? 'light';
     const spawnSpec = (0, dispatch_1.resolveTierSpawn)(dispatch, tier, memberIssue);
     const cmd = spawnSpec.cmd;
     // #677: the prompt carries THIS member's own worktree and the integration
     // branch it was cut from — `member-cycle`'s `worktree`/`integration_branch`
     // inputs. The shared batch worktree is no longer a member surface at all.
-    const prompt = (0, dispatch_1.buildMemberPrompt)(dispatch.memberPrompt, memberIssue, batchId, member.worktree, batch.branch);
+    const prompt = (0, dispatch_1.buildMemberPrompt)(dispatch.memberPrompt, memberIssue, batchId, member.worktree, batch.branch, review);
     const logFile = (0, dispatch_1.batchMemberLogPath)(deps.store.runsDir, batchId, batch.executing_member, memberIssue);
     // #629: captured BEFORE spawning, mirroring `engine.ts`'s own
     // `spawnAndRecord` — the log is per-role and append-mode, so the size at
@@ -966,6 +977,7 @@ function spawnMember(deps, dispatch, state, slot, batchId, now, result, member)
     deps.journal.append((0, journal_1.unitEvent)('spawned', unit(batchId), {
         pid,
         tier,
+        ...(review === 'full' ? { review } : {}),
         slot: slot.id,
         issue: memberIssue,
         ...(0, dispatch_1.journalCmdModelFields)(spawnSpec),
@@ -1342,7 +1354,7 @@ function runValidate(deps, config, dispatch, batchId, now, result) {
     const batch = (0, state_1.findBatch)(state, batchId);
     if (!batch || batch.worktree === null)
         return;
-    const suite = safeSuite(deps, batchId, batch.worktree);
+    const suite = safeSuite(deps, batch, batch.worktree);
     const rDeps = recoveryDeps(deps, config, batch, now);
     const poster = (0, recovery_1.createExecMilestonePoster)(deps.exec, { repoDir: deps.repoDir });
     if (suite.ok) {
@@ -1984,7 +1996,10 @@ function recordMemberRunLog(deps, dispatch, state, batchId, batch, memberIssue,
         });
         return null;
     }
-    const tier = (0, state_1.findEntry)(state, memberIssue)?.tier ?? 'mid';
+    // Same effective tier `spawnMember` dispatched at (#771), so the run log
+    // records the model that actually ran.
+    const memberEntry = (0, state_1.findEntry)(state, memberIssue);
+    const tier = memberEntry ? (0, dispatch_1.memberDispatchTier)(memberEntry) : 'mid';
     const { cmd, model } = (0, dispatch_1.resolveTierSpawn)(dispatch, tier, memberIssue);
     const logFile = (0, dispatch_1.batchMemberLogPath)(deps.store.runsDir, batchId, batch.executing_member, memberIssue);
     // #629: fenced to THIS dispatch's slice (`log_offset_at_spawn`, stamped by
@@ -2333,7 +2348,7 @@ function reconcileFixSlot(deps, config, batchId, slot, now, result) {
     // packages/sched/README.md's dispatch-health section.
     handleDeadDispatchApiError(deps, batchId, (0, dispatch_1.batchFixLogPath)(deps.store.runsDir, batchId, offenderRecord.issue), slot.log_offset_at_spawn ?? 0, now, { release: false });
     deps.store.withLock((s) => ({ state: releaseSlot(s, batchId, now), result: undefined }));
-    const suite = safeSuite(deps, batchId, batch.worktree);
+    const suite = safeSuite(deps, batch, batch.worktree);
     const rDeps = recoveryDeps(deps, config, batch, now);
     const { state: resolved } = (0, recovery_1.resolveFixAttempt)(deps.store.load(), batchId, offenderRecord.issue, suite.ok ? 'green' : 'red', rDeps);
     deps.store.withLock((s) => ({