@ai-dossier/sched 0.46.0 → 0.47.0

package/README.md

@@ -32,7 +32,7 @@ ai-dossier sched status           # queue (+pr/cleanup), engine lease, slots, ba
 ai-dossier sched pause            # prevent every new agent process; live units keep running
 ai-dossier sched resume
 ai-dossier sched stop --issue 42  # terminate one full-cycle agent and record it stopped (no recovery)
-ai-dossier sched stop --batch b1  # terminate the batch process and stop unfinished members
+ai-dossier sched stop --batch b1  # terminate every batch agent (incl. parallel members) and stop unfinished members
 ai-dossier sched abandon --issue 42 --reason "operator abort"
 ai-dossier sched abandon --batch b1   # dissolve; members requeue as full-cycle
 ai-dossier sched stats --issues 4..9  # per-issue tokens/cost from ~/.dossier/runs.jsonl (#524)
@@ -53,8 +53,8 @@ directory directly, reconstructing costs from the raw dispatch logs rather than
 takeovers; it does not terminate agents that are already running. `stop --issue <n>` is the
 single-command stop path: it PID-start-safely terminates that issue's live agent, releases its
 slot, and records a terminal `stopped` outcome that will not recover or escalate. An active
-slot-mode member must instead be stopped through `stop --batch <id>`, which terminates the batch
-process and atomically stops the batch and its unfinished members. `abandon` only records failure
+slot-mode member must instead be stopped through `stop --batch <id>`, which terminates every agent
+the batch holds (its own slot plus one per running parallel member, `batch:<id>#<issue>`) and atomically stops the batch and its unfinished members. `abandon` only records failure
 and releases the slot; it intentionally does not terminate its process.
 
 Since #507, `enqueue` additionally reads each candidate issue's live GitHub labels (one
@@ -688,9 +688,15 @@ member got the light, relevance-scoped review). `review=full` is how a risk-floo
   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.
+- **Prompt** — `buildMemberPrompt` substitutes a `{review}` placeholder.
+  `DEFAULT_MEMBER_PROMPT_TEMPLATE` places `{review}` itself (#804), so both `light` and `full`
+  members see their level and are told to run member-cycle Step 4b's self-review at it before
+  handover (`full` → the full review-issue tier incl. Security; `light` → at least
+  Conformance), then post `review done` naming the agents that ran — never `agents_done=0`.
+  An operator `member_prompt` template without `{review}` gets `FULL_REVIEW_MEMBER_DIRECTIVE`
+  appended for a `review=full` member; a `light` member's prompt from such a template is
+  byte-identical to before. A member's `review partial` (an agent could not finish) is a
+  terminal hand-back like `blocked` (`reason=review-partial:<agents_pending>`).
 - **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.
 
@@ -748,7 +754,7 @@ only when `batchExec`/`runBatchSuite` are both configured on `EngineDeps` — dr
 `batch:<id>` unit through:
 
 ```
-ready → executing(member i/N) ⟲ → validating → reviewing → shipping
+ready → executing(serial: member i/N ⟲ | parallel: members ∥, ordered landing) → validating → reviewing → shipping
   → awaiting-merge → merged → deployed → reported → done
 failure rails: executing → dissolving (a member self-reports blocked)
                validating → attributing → (fixing | evicting) → validating → dissolving
@@ -768,7 +774,25 @@ failure rails: executing → dissolving (a member self-reports blocked)
   `warm_commands` even in repos that never use the pool for anything else).
   This shared tree is the batch's INTEGRATION branch: members branch off it, and
   the tail work (aggregate suite, review, ship) runs in it (#677).
-- **Members run `member-cycle` serially, one fresh agent at a time — each in its OWN
+- **Members run in PARALLEL by default (#809)** — each member's `member-cycle` agent holds
+  its OWN slot (`batch:<id>#<issue>`, bounded by config `member_parallelism`, default
+  `max_slots`, and free capacity), in its own worktree cut off the integration branch, so
+  a batch's member phase costs ≈ max(member), not sum(member). Nothing lands until a
+  member-order prefix is verified: each verified member is rebased onto the integration
+  TIP in its own worktree, its pushed branch refreshed (`--force-with-lease`), then
+  `merge --ff-only`-landed — the same linear, `(#<issue>)`-trailed history (and
+  `memberRanges` attribution) as serial landing. A member whose diff no longer rebases
+  onto the members landed before it is evicted (`landing-conflict`, requeued full-cycle,
+  its pushed branch kept) — batch-integrate's eviction verdict; a parent repair agent (#813) for
+  that conflict is a follow-up. A gate-inconclusive member still lands and keeps its
+  tree; once every member resolves the batch blocks on it exactly as in serial mode
+  (`sched resume --batch` works unchanged). The mode is decided ONCE at the
+  `ready → executing` claim and persisted (`BatchEntry.member_dispatch`, `member_runs`
+  per member; schema 1.23.0): **serial** when `member_parallelism` is 1, the batch has
+  an eviction group, or a member `deps` on another member; a batch already past `ready`
+  with no recorded mode (claimed by a pre-1.23.0 engine) stays serial across the upgrade.
+  Dissolve, `sched stop --batch` and `abandon` stop/release every member slot.
+- **Serial mode: members run `member-cycle` one fresh agent at a time — each in its OWN
   worktree** on its OWN branch `batch/<id>-m<n>-<issue>` (#677, RFC-0001 §J.3), cut off
   the integration branch, warmed, and pushed before the agent spawns; the agent never
   creates either. When the member's incremental gate passes, the scheduler LANDS the
@@ -795,8 +819,9 @@ failure rails: executing → dissolving (a member self-reports blocked)
   gate later to resolve the block once the capability is fixed, and applies the same
   evidence bar on the recheck — a still-unevidenced `task-failed` stays blocked rather
   than evicting on a resume.
-- **The batch's single slot is claimed FRESH for each live step** (a member, the tail
-  agent, the report agent, a bounded fix agent) — never held across a wait. The aggregate
+- **The batch's own slot is claimed FRESH for each live step** (a serial member, the tail
+  agent, the report agent, a bounded fix agent) — never held across a wait; parallel
+  members hold their own member slots instead. The aggregate
   suite itself runs with NO slot claimed at all (deterministic engine work, not an LLM
   step).
 - **Two failure rails.** A member that never went green evicts directly (nothing to
@@ -1022,6 +1047,8 @@ import {
   duplicateEvictionDetail, // the one eviction-duplicate wording, shared by both rails
   isPreservedMember,     // the single definition of "already green"
   createBatch,           // the single BatchEntry constructor
+  batchMemberUnit,       // #809: a parallel member's slot unit, `batch:<id>#<issue>`
+  slotsForBatch,         // #809: every live slot a batch holds (its own + member slots)
   type RecoveryDeps,     // inject exec/repoDir/journal/milestone-poster/suite-runner/clock
   type SuiteRunner,      // re-runs the aggregate suite after a revert or rebase
   type BatchMilestonePoster, // batch-milestone sink (createExecMilestonePoster is default)
@@ -1102,7 +1129,8 @@ telemetry" below.
 │                  # pr_poll_interval_ms, dispatch (incl. report_prompt,
 │                  # phase_stall_timeout_ms, fence_takeover_timeout_ms, tiers — #527,
 │                  # suite_command — #562, disallowed_tools — #591), auto_upgrade — #537,
-│                  # dissolve_policy — #563
+│                  # dissolve_policy — #563, default_batch_priority — #565,
+│                  # max_full_review_members — #771, member_parallelism — #809
 ├── events.jsonl   # append-only event journal (the operator's flight recorder)
 ├── runs/          # per-unit agent output logs (issue-<n>.log)
 └── .sched-lock/   # cross-process directory mutex (pid; stolen from dead holders)
@@ -1209,8 +1237,7 @@ after-the-fact recovery, not a missing-data bug.
   queue data.
 - **Schema**: state/config files from #460 (schema 1.0.0), #464 (1.1.0), #468 (1.2.0),
   #472 (1.3.0), #500 (1.4.0), #505 (1.5.0), #504 (1.6.0), #523 (1.7.0) and #524 (1.8.0)
-  load and migrate to 1.14.0 automatically (slot `branch`/`last_head`/`pid_start`, slot `role` (inferred from the
-  load and migrate to 1.13.0 automatically (slot `branch`/`last_head`/`pid_start`, slot `role` (inferred from the
+  load and migrate to the current schema (1.23.0) automatically (slot `branch`/`last_head`/`pid_start`, slot `role` (inferred from the
   unit's queue entry, with the persisted `phase` as a fallback — #500), entry
   `pr`/`cleanup`/`failure_evidence`, batch `anchor`/`branch`/`run_id`/`eviction_groups`/
   `evictions`/`fix_attempts`/`rebase_attempts`, state-level `last_pr_poll_at` backfill to
@@ -1222,7 +1249,8 @@ after-the-fact recovery, not a missing-data bug.
   `stale_milestone_ignored_for` backfill to `null` — #610, and state-level
   `consecutive_dispatch_api_errors`/`dispatch_pause_reset_at` backfill to `0`/`null` —
   #629; batch `anchor_closed_at` and `anchor_close_failed_reason`/`_since`/`_ticks`
-  backfill to `null`/`null`/`null`/`0` — #768, schema 1.22.0).
+  backfill to `null`/`null`/`null`/`0` — #768, schema 1.22.0; batch `member_dispatch`/`member_runs`
+  backfill to `null`/`[]` — #809, schema 1.23.0 — a null mode past `ready` runs serially).
 - **`max_slots`** bounds live units (`assigned | running | recovering`); dependency
   edges gate readiness — an issue with an unmerged dependency, and a batch behind an
   unmerged batch, are never runnable.

package/dist/batch-dispatch.d.ts

@@ -11,7 +11,8 @@
  * at batch-phase granularity instead of per-issue-phase granularity:
  *
  * ```
- * ready → executing(member i/N) ⟲ → validating → reviewing → shipping
+ * ready → executing(serial: member i/N ⟲ | parallel: members ∥ + ordered landing)
+ *   → validating → reviewing → shipping
  *   → awaiting-merge → merged → deployed → reported → done
  * failure rails (RFC F.2/F.8/F.9):
  *   executing → dissolving  (a member self-reports blocked, RFC F.1)
@@ -67,7 +68,7 @@ import type { SchedStore } from './persist';
 import type { ExecFn } from './project';
 import { type BatchSuiteContext, type SuiteResult } from './recovery';
 import { type FsExists } from './teardown';
-import type { BatchEntry, CapabilityGateResult, SchedConfig } from './types';
+import type { BatchEntry, CapabilityGateResult, MemberDispatchMode, SchedConfig, SchedState } from './types';
 export type { CapOutcome } from './types';
 /** Everything batch dispatch needs from the outside world. */
 export interface BatchDispatchDeps {
@@ -230,6 +231,20 @@ export declare function resumeBlockedGate(deps: BatchDispatchDeps, config: Sched
     blockedBy?: 'inconclusive' | 'unevidenced-failure';
     detail?: string;
 };
+/**
+ * How a batch should dispatch its members (#809) — decided once, at the
+ * `ready → executing` claim, and persisted on `BatchEntry.member_dispatch`.
+ *
+ * Serial when the operator set `member_parallelism: 1`, or the batch itself
+ * says its members are not independent: an eviction group (members that must
+ * revert together — e.g. one built on another's API) or a member whose `deps`
+ * names another member (it must see that member's landed work before it is
+ * cut). Parallel otherwise — the batch-cycle skill's and prep dossier's
+ * design (RFC-0001 §J.3: each member works off the integration branch and
+ * sees no one else's changes until the parent lands them).
+ */
+/** @internal exported for tests. */
+export declare function memberDispatchModeFor(state: SchedState, batch: BatchEntry, config: SchedConfig): MemberDispatchMode;
 /**
  * One batch reconcile+refill pass. Called from `engine.ts`'s `tick()` after
  * the issue-level pass — this pass never claims a slot the issue pass already

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;AA8BH,OAAO,EAaL,KAAK,gBAAgB,EAGrB,KAAK,SAAS,EACf,MAAM,YAAY,CAAC;AAEpB,OAAO,EAEL,KAAK,WAAW,EAUjB,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;AAqBpB,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;IACjB;;;;;;OAMG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;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;AAkxCD;;;;;;;;;;;;;GAaG;AACH,wBAAgB,YAAY,CAC1B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,GACzB,eAAe,CAiQjB"}
+{"version":3,"file":"batch-dispatch.d.ts","sourceRoot":"","sources":["../src/batch-dispatch.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8DG;AA8BH,OAAO,EAaL,KAAK,gBAAgB,EAGrB,KAAK,SAAS,EACf,MAAM,YAAY,CAAC;AAEpB,OAAO,EAEL,KAAK,WAAW,EAWjB,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;AA0BpB,OAAO,EAAE,KAAK,QAAQ,EAA2D,MAAM,YAAY,CAAC;AACpG,OAAO,KAAK,EAEV,UAAU,EAEV,oBAAoB,EAKpB,kBAAkB,EAIlB,WAAW,EACX,UAAU,EAEX,MAAM,SAAS,CAAC;AAcjB,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;IACjB;;;;;;OAMG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;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;AAwiBD;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,MAAM,CAE3F;AA23CD;;;;;;;;;;;;;;;;;;;;;;;;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;AA2TD;;;;;;;;;;;;;;;;;;;;;;;;;;;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;AAi0CD;;;;;;;;;;;GAWG;AACH,oCAAoC;AACpC,wBAAgB,qBAAqB,CACnC,KAAK,EAAE,UAAU,EACjB,KAAK,EAAE,UAAU,EACjB,MAAM,EAAE,WAAW,GAClB,kBAAkB,CAQpB;AAizBD;;;;;;;;;;;;;GAaG;AACH,wBAAgB,YAAY,CAC1B,IAAI,EAAE,iBAAiB,EACvB,MAAM,EAAE,WAAW,EACnB,QAAQ,EAAE,gBAAgB,GACzB,eAAe,CA6RjB"}