mandrel 1.94.0 → 2.0.0

package/.agents/workflows/helpers/deliver-epic.md

@@ -1,998 +0,0 @@
----
-description: >-
-  Drive an Epic from `agent::ready` to a merged pull request against `main`.
-  The ten-phase flow runs the wave loop, close-validation, the Epic-close
-  lens-roster resolve, the code-review pass (which walks the cumulative diff
-  once with the lens roster folded in), retro, finalize, watch-and-iterate,
-  conditional auto-merge,
-  and local branch cleanup. When the run is end-to-end clean (zero manual
-  interventions, zero 🔴/🟠 review findings, compact retro) the PR
-  auto-merges via `gh pr merge --squash --delete-branch`; otherwise the
-  workflow falls back to the operator-merges-button path so a human
-  inspects the surface area.
----
-
-# helpers/deliver-epic — Epic delivery path (invoked by /deliver)
-
-> **Runtime core.** This file is the always-ingested Epic-delivery path:
-> phase flow, commands, gate contracts, and return shapes. The recovery
-> procedures, historical rationale, and troubleshooting detail live in the
-> sibling [`deliver-epic-reference.md`](deliver-epic-reference.md); each
-> moved procedure keeps a one-line pointer at its trigger point below. The
-> reference is not projected to `.claude/commands/` — it is consulted on
-> demand.
-
-## Overview
-
-This helper is the **Epic delivery path** behind `/deliver` — the router
-delegates to it once per Epic ID, either as the sole route (single-Epic
-input) or as one **Epic segment** of the sequential segment plan `/deliver`
-composes over mixed Epic / standalone-Story input (Epic segments run in
-input order, after the standalone segment; see
-[`deliver.md`](../deliver.md)). Each invocation opens a PR against `main`
-and auto-merges when every signal certifies a clean run; otherwise it falls
-back to the operator-merges-button path.
-
-```text
-/deliver <epicId>
-  → Phase 1 — prepare              (epic-deliver-prepare.js)
-  → Phase 2 — ready-set loop       (wave-tick.js → dispatch ready set → observe → re-tick)
-  → Phase 3 — close-validation     (lint + test + ratchets on epic/<id>)
-  → Phase 4 — epic-close lens roster (helpers/epic-audit.md — resolve the slim cumulative+global+risk-routed roster)
-  → Phase 5 — code-review          (helpers/code-review.md scope: epic — walks the cumulative diff once, lens roster folded in)
-  → Phase 6 — retro                (.agents/scripts/lib/orchestration/retro-runner.js)
-  → Phase 6.5 — integration gate   (whole-product navigability + journey suite; @pending ≠ green for surface-adding Epics — blocks finalize)
-  → Phase 7 — finalize             (lifecycle-emit → epic.close.end → open PR to main)
-  → Phase 8 — watch-and-iterate    (poll `gh pr checks`; fix locally until green)
-  → Phase 8.5 — auto-merge gate    (lifecycle-emit → epic.automerge.start)
-  → Phase 9 — cleanup              (BranchCleaner + Cleaner lifecycle listeners on epic.cleanup.start / epic.merge.armed; fire via lifecycle-emit → epic.merge.armed)
-```
-
-The argument is always a single Epic ID (`type::epic`) — multi-Epic or
-mixed input is segmented by the `/deliver` router before this helper runs.
-Story IDs go to
-[`helpers/deliver-stories`](deliver-stories.md) (standalone) or the
-[`helpers/epic-deliver-story`](epic-deliver-story.md) helper
-(Epic-attached, invoked by this workflow's fan-out); Tasks are not directly
-executable.
-Story dispatch is in-session via the Agent tool — no subprocess is
-spawned.
-
----
-
-## Arguments
-
-```text
-/deliver <epicId> [--skip-epic-audit] [--skip-code-review] [--skip-retro] [--full-retro] [--skip-integration-gate]
-```
-
-- `epicId` — must carry `type::epic`. Otherwise STOP and tell the operator
-  to use `/deliver <id>` (standalone Story) or open the parent Epic.
-- `--skip-epic-audit` — skip Phase 4 (log the override). Use only when the
-  change-set audits are known to be irrelevant (e.g., docs-only Epic).
-- `--skip-code-review` — skip Phase 5 (log the override).
-- `--skip-retro` — skip Phase 6 (use sparingly).
-- `--full-retro` — force the six-section retro regardless of manifest
-  cleanliness. `--skip-retro` wins over `--full-retro`.
-- `--skip-integration-gate` — skip Phase 6.5 (log the override). The
-  explicit operator override for the post-wave integration gate,
-  consistent with `--skip-epic-audit`. Skipping the gate is recorded as a
-  manual intervention and disqualifies auto-merge, exactly like the other
-  `--skip-*` overrides.
-
-Every other runtime modifier is sourced from the Epic's labels or from
-`delivery.deliverRunner` in `.agentrc.json`.
-
-- `--yes` — the top-level `/deliver` unattended flag (suppresses the
-  operator confirmation gate; see [`deliver.md`](../deliver.md)). This
-  helper reads it as the **headless signal** for Phase 8.5's must-land
-  terminal step (Story #4427): when `/deliver` was invoked with `--yes`,
-  thread `--headless true` into the Phase 8.5
-  `epic.automerge.start` `lifecycle-emit.js` call (omit the flag, or
-  pass `--headless false`, for an attended run). The runtime reads this
-  as an explicit `headless` input threaded through
-  `runLifecycleEmit` → `buildDefaultListenerChain` →
-  `MergeWatcher({ headless })` — never an ambient/undefined global.
-  Attended-mode behavior (no `--headless` flag) is byte-for-byte
-  unchanged: `MergeWatcher` defaults `headless` to `false`.
-
----
-
-## Contract
-
-- **Idempotent by checkpoint.** Re-runs resume from `epic-run-state`.
-- **Single pause point.** Only `agent::blocked` halts execution. No
-  clarifying questions — if stuck, flip to `agent::blocked`, post a
-  friction comment, park.
-- **Flat Story dispatch by design.** Host LLM fans out per-Story Agent
-  calls directly — with `subagent_type: story-worker` when
-  `delivery.routing.roleScopedAgents` is on (the default; § 2b), else
-  `general-purpose`. Keeping Story
-  dispatch flat — the host owns the single fan-out level — is a
-  **design choice**, not a harness constraint: the wave aggregator, idle
-  watchdog, and merge-lock all assume one host-owned dispatch level. As of
-  Claude Code 2.1.202 a level-1 sub-agent **does** carry the `Agent` tool
-  and can nest further (verified depth 2, announced max depth 5; see
-  [#2870](https://github.com/dsj1984/mandrel/issues/2870)), so a Story
-  worker may itself fan out for its own sub-work within that depth budget —
-  the Epic wave loop nonetheless stays flat by choice, not because nesting
-  is unavailable.
-- **Operator-merges-PR exit.** Phase 7 opens the PR; the workflow
-  never merges to `main` itself. Phase 8.5 may fire auto-merge when
-  every signal is clean.
-- **Lifecycle bus is the runner model.** Phase transitions, ticket
-  state flips, structured comments, and notifications are emitted as
-  typed events on the in-session lifecycle bus; a fixed roster of
-  listeners performs the side effects. Phase 7, 8.5, and 9 each fire
-  exactly one lifecycle event via the generic
-  [`lifecycle-emit.js`](../../scripts/lifecycle-emit.js) CLI
-  (`--event epic.close.end` / `--event epic.automerge.start` /
-  `--event epic.merge.armed`); the matching listener chain runs the
-  bus-driven side effects (acceptance reconcile, automerge-armer,
-  branch cleanup). The append-only NDJSON ledger at
-  `temp/epic-<id>/lifecycle.ndjson` is the resume target. See
-  [`docs/LIFECYCLE.md`](../../../docs/LIFECYCLE.md) for the bus
-  contract, event taxonomy, ledger format, and listener model.
-
-> **Hierarchy.** `/deliver` operates over the 2-tier hierarchy
-> (Epic → Story). The fan-out is one `Agent` tool call per
-> Story per wave (§ 2b); Story branches merge into `epic/<id>` with
-> `--no-ff` via `story-close.js`; the close-validation chain
-> (Phase 3), epic-audit, code-review, retro, finalize, and auto-merge
-> gates all operate on Story-level units.
-> [`helpers/epic-deliver-story`](epic-deliver-story.md) runs a
-> single Story-implementation phase per Story against the Story's
-> inline `acceptance[]` / `verify[]` fields. See
-> [`.agents/instructions.md` § 5.D](../../instructions.md) and
-> [`.agents/docs/SDLC.md` § Ticket hierarchy](../../docs/SDLC.md) for the full
-> contract.
-
----
-
-## Phase 1 — Prepare the Epic run
-
-### Phase 1 prelude — Delivery preflight (Story #2899 / F13)
-
-Before `epic-deliver-prepare.js` seeds the checkpoint, run
-`epic-deliver-preflight.js` so the operator (and any reviewer reading the
-Epic ticket) sees the estimated Story count, install cost, dependency depth,
-GitHub API request volume, and Claude Max quota burn for the run that is
-about to fan out. **Preflight always runs before Story fan-out.**
-
-```bash
-node .agents/scripts/epic-deliver-preflight.js --epic <epicId> --post
-```
-
-The CLI upserts a `delivery-preflight` structured comment on the Epic
-(idempotent across re-runs) and prints a JSON envelope on stdout with
-the canonical metric keys `storyCount`, `installCostSeconds`,
-`dependencyDepth` (the longest dependency chain — the ready-set wall-clock
-floor, replacing the retired wave count), `githubApiRequests`,
-`claudeQuotaTokens`, plus a `breaches` array describing any
-`delivery.preflight.max*` thresholds the estimate exceeds.
-
-**Breach handling.** When `breaches` is non-empty, the workflow MUST
-flip the Epic to `agent::blocked`, surface the envelope in chat for the
-operator, and halt before Phase 1's `epic-deliver-prepare.js` call.
-Resume after the operator unblocks (raising the threshold in
-`.agentrc.json`, splitting the Epic, or accepting the cost) by re-running
-`/deliver <epicId>` — the preflight is idempotent and the second
-run upserts the same comment in place.
-
-Threshold defaults live in `delivery.preflight.*` in `.agentrc.json`
-(all keys default to "no cap" — the gate is opt-in until an operator
-configures `maxStories` etc.).
-
-**Remote evidence — land or block (issue #4483).** The envelope also
-carries `remoteVerified` + `remoteProbe` (deterministic probes:
-`git remote get-url origin`, bounded `git ls-remote origin HEAD`). When
-`remoteVerified` is `false`, flip the Epic to `agent::blocked`, post a
-friction comment quoting `remoteProbe.detail`, and halt — the same
-explicit-block shape as #4425/#4480. NEVER fall back to executing Stories
-inline in this session or committing the delivery to local `main`; the
-worktree/branch/PR path below is the only sanctioned mechanism. Phase 7's
-finalize additionally refuses with a `delivery-branch-missing-on-origin`
-blocker when `epic/<epicId>` never reached origin.
-
-### Phase 1 main — Seed the wave plan
-
-```bash
-node .agents/scripts/epic-deliver-prepare.js --epic <epicId> [--steal] [--as <handle>]
-```
-
-Validates `type::epic`, enumerates `type::story` descendants, parses
-`blocked by #N` plus explicit `dependencies`, computes the dependency DAG
-(to enumerate the open Story set), and upserts the `epic-run-state`
-checkpoint in the per-Story-status shape (a flat `stories` map seeded at
-`pending`, plus the global `concurrencyCap`). Treat the printed JSON as
-`state`: `{ epicId, storyCount, concurrencyCap, stories, checkpointInitializedAt, docsDigestPath }`.
-`stories` is the flat dispatch hint (`{ storyId, worktree, title,
-checklistPath }` per open Story); the ready-set `tick` (Phase 2) decides
-which to dispatch on each beat. `checklistPath` is the repo-relative path to
-that Story's footprint-matched local-lens authoring checklist
-(`temp/epic-<epicId>/checklists/story-<storyId>.md`, Story #4410) — thread it
-into that child's prompt (§ 2b, item 7); it is `null` when the Story's
-predicted footprint matched no local lens. `docsDigestPath` is the
-repo-relative path to the per-Epic docs digest
-(`temp/epic-<epicId>/docs-digest.md`) that prepare writes from
-`project.docsContextFiles` — thread it into every child prompt (§ 2b, item 6).
-It is `null` when the project configured no `docsContextFiles` (no digest is
-written). Flip the Epic to `agent::executing` (idempotent) after the CLI
-returns.
-
-**No spec-ticket linkage to resolve (Story #4324).** The Tech Spec lives
-as managed sections of the Epic body itself — there is no separate
-Tech-Spec issue id in the envelope and no `--tech-spec` flag to thread
-into the per-Story `story-init.js` invocations. Story agents receive the
-Tech Spec via context hydration, which embeds the Epic body (with the
-`## Acceptance Table` section stripped) directly into each Story prompt.
-
-> **Preflight guards + acceptance-table start gate.** Before the snapshot
-> phase runs — and before any worktree is created — prepare runs two
-> **fail-closed** guards (checkout safety + Epic lease), then the snapshot
-> phase asserts the Epic carries an `acceptance::n-a` waiver or a
-> `## Acceptance Table` section. Both throw on failure. See
-> [`deliver-epic-reference.md` § Phase 1 — Preflight guards](deliver-epic-reference.md#phase-1--preflight-guards-story-3482--f-workflow-guards)
-> for the remediation detail (dirty-tree recovery, `--steal`, waiver
-> options).
-
----
-
-## Phase 2 — Ready-set loop
-
-The scheduler lives in
-[`lib/wave-runner/tick.js`](../../scripts/lib/wave-runner/tick.js) — a thin
-**Epic adapter over the ready-set core**
-([`lib/wave-runner/ready-set.js`](../../scripts/lib/wave-runner/ready-set.js)).
-One stateless `tick({ epic })` call re-derives readiness from the **live**
-Story bodies + labels on every beat and returns one `WaveTickResult`
-describing the next action. There is **no wave barrier** (Story #4155): a
-Story whose own dependencies are all done is dispatched the instant a slot is
-free under the GLOBAL in-flight cap, even while an unrelated sibling Story is
-still `agent::executing`. The loop is simply:
-
-```text
-tick → dispatch the ready set → observe → re-tick → … → epic-complete
-```
-
-The slash command's job each beat is to call `tick()` via its CLI shim,
-dispatch the Stories in `nextAction.stories` via the Agent tool, record each
-returned Story's terminal status, and re-tick until terminal. There is no
-`record-wave` / `currentWave` step — the checkpoint carries only a flat
-per-Story status map (for resume + the operator rollup) and the global cap.
-
-### 2.0. Open the Epic PR as a draft at wave 1 (Story #4359)
-
-When `delivery.ci.earlyPr` is on (the default), open the Epic PR as a
-**draft** once, before the first `tick`, so every subsequent per-wave push
-to `epic/<epicId>` runs CI attributed to its own wave. (CI is keyed on the PR
-ref with `cancel-in-progress`, so each new wave push **supersedes** the prior
-wave's in-flight run rather than queuing behind it — the latest wave always
-gets the verdict, and CI-minute use stays bounded; intermediate wave runs are
-cancelled, not completed.) Resolve the flag through the
-[`getCiDelivery`](../../scripts/lib/config/ci.js) accessor (default `true`);
-do not read `delivery.ci.earlyPr` directly.
-
-> **This step is host-LLM-driven, with no runtime enforcement seam** — unlike
-> the Phase 7 ready-flip, which the `Finalizer` listener wires in
-> deterministically (`finalizer.js` resolves `earlyPr` and calls
-> `markPrReady`). The asymmetry is intentional and safe: if this wave-1
-> draft-open is skipped, the `earlyPr`-on Phase 7 `markPrReady` call degrades
-> to a no-op on the PR that finalize opens at close time (`gh pr ready` is a
-> no-op on an already-ready PR), so the merge gate is never stranded — the run
-> only loses the per-wave CI attribution this step buys.
-
-- **`earlyPr` on** — call
-  [`openOrLocatePr`](../../scripts/lib/orchestration/finalize/open-or-locate-pr.js)
-  with `{ epicId, headBranch: 'epic/<epicId>', baseBranch: 'main', draft: true }`.
-  The helper probes for an existing open PR first, so this is idempotent — a
-  resumed `/deliver` run re-locates the same draft and opens no duplicate.
-  Phase 7 later flips this draft to ready-for-review (it does **not**
-  re-create the PR).
-- **`earlyPr` off** — skip this step entirely. No draft is opened at wave 1;
-  Phase 7 opens the PR at close time on the pre-Story timing.
-
-The draft carries the same title/body contract Phase 7 uses
-(`feat: Epic #<epicId>` / `Closes #<epicId>`), so no title/body reconciliation
-is needed when it is marked ready.
-
-### 2a. Tick — plan the next action
-
-```bash
-node .agents/scripts/wave-tick.js --epic <epicId>
-```
-
-Stdout is one `WaveTickResult` envelope:
-
-```json
-{
-  "nextAction":
-      { "kind": "dispatch", "stories": [{ "id": <n>, "title": "…" }, ...], "in-flight": [<storyId>, ...] }
-    | { "kind": "observe",  "waitingOn": [<storyId>, ...], "in-flight": [<storyId>, ...] }
-    | { "kind": "halt", "reason": "dependency-cycle" | "unsatisfiable-dependency", "stuckStories": [<storyId>, ...], "cycle"?: [<storyId>, ...], "in-flight": [<storyId>, ...] }
-    | { "kind": "epic-complete", "in-flight": [<storyId>, ...] },
-  "blockedStories": [{ "storyId": <n>, "reason": "…", "detail"?: "…" }, ...],
-  "gateFailures":   [{ "storyId": <n>, "gate": "…", "detail"?: "…" }, ...],
-  "readyCount":     <n>,
-  "inFlight":       [<storyId>, ...]
-}
-```
-
-`nextAction.stories` is the **ready set** for this beat — the
-dependency-satisfied, overlap-free subset of open Stories, capped at
-`globalCap − inFlight`. The CLI is a planner: it dispatches nothing and
-persists nothing. It emits only the two wave-window forensics signals that
-have a live consumer — `wave-start` (on the run's first dispatch) and
-`wave-complete` (when the run finishes), which the perf-aggregator brackets
-into the `waveParallelism` report (and `wave-start` anchors span-tree Story
-spans). The [`signals` helper](signals.md)
-(`node .agents/scripts/signals-view.js`) renders the forensics signals in the
-span-tree view.
-
-> **Old-shape checkpoint → fail-closed.** A pre-ready-set
-> (`plan` / `currentWave` / `totalWaves`) checkpoint makes the tick refuse
-> to run and throw; re-seed via `epic-deliver-prepare.js`. See
-> [`deliver-epic-reference.md` § Fail-closed on an old-shape checkpoint](deliver-epic-reference.md#fail-closed-on-an-old-shape-checkpoint).
-
-### 2b. Dispatch — fan out per-Story Agent calls
-
-*You* (the LLM running this skill) are the dispatcher; you never invoke
-`helpers/epic-deliver-story` yourself. Emit **one `Agent` tool call per
-Story** in `nextAction.stories` (even when `length === 1` — the
-parent-child boundary keeps the return-parser uniform). The *children*
-run [`helpers/epic-deliver-story`](epic-deliver-story.md).
-
-**Sub-agent type (Epic #4478, M7-B).** When
-`delivery.routing.roleScopedAgents` is enabled (the **default**), each Story
-child is dispatched with `subagent_type: story-worker` — it boots on the
-role-scoped [`story-worker`](../../agents/story-worker.md) context (its own
-system prompt, no `CLAUDE.md` @-closure) that carries every load-bearing
-delivery MUST standalone, so the spawn stops re-paying the always-loaded
-context. When the kill-switch is **off**
-(`delivery.routing.roleScopedAgents: false`), fall back to
-`subagent_type: general-purpose` — the instant, code-rollback-free revert and
-the escape for hosts that ignore `.claude/agents/`.
-
-Emit **one assistant turn** with **N parallel `Agent` calls** where
-`N === nextAction.stories.length` (the ready set is already capped at
-`globalCap − inFlight` by the tick, so it never exceeds available slots).
-Dispatch the ready set as background calls (`run_in_background: true`) and,
-as each child returns, record it (§ 2c) and **re-tick** (§ 2a) to pull the
-next ready set — never wait for the whole set before refilling.
-
-> **Throughput + capability tuning.** The default `concurrencyCap` of 3 is
-> a deliberate operator-tuning knob (raise
-> `delivery.deliverRunner.concurrencyCap`), and the optional per-call
-> `model:` escape hatch lets mechanical Stories run on a cheaper capability.
-> See [`deliver-epic-reference.md` § Throughput tradeoff](deliver-epic-reference.md#throughput-tradeoff)
-> and [§ Sub-agent dispatch capability](deliver-epic-reference.md#sub-agent-dispatch-capability).
-
-**Ledger the dispatch BEFORE the Agent call.** Immediately before each
-per-Story `Agent` tool call (one shell-out per Story, every attempt —
-including retries from a refill), invoke
-[`lifecycle-emit-story-dispatch.js`](../../scripts/lifecycle-emit-story-dispatch.js)
-so the lifecycle ledger durably records the dispatch attempt. The
-emit must happen **before** the Agent call fires — never after — so
-that a host-process crash mid-Agent leaves a `story.dispatch.start`
-record that `wave-tick.js` (see § 2a) excludes from the next beat's ready
-set and surfaces under `nextAction['in-flight']`:
-
-```bash
-node .agents/scripts/lifecycle-emit-story-dispatch.js \
-  --epic <epicId> --story <storyId> \
-  --wave 0 --attempt <attempt>
-```
-
-Pass `--wave 0` — the ready-set runtime has a single continuous front, so
-the ledger's `waveIndex` is a fixed `0` (it is metadata for the start/end
-pairing math, not a scheduling input). `<attempt>` starts at 1 for the
-Story's first dispatch and increments on each retry/refill. The CLI appends
-exactly one NDJSON line to `temp/epic-<epicId>/lifecycle.ndjson`; the
-matching `story.dispatch.end` record is appended later by
-`epic-execute-record-wave.js` (via `emit-story-dispatch-end.js`, Story #3900)
-after the Agent return is recorded in § 2c.
-
-Each Agent call's prompt must (1) name the Story + Epic ids, (2)
-instruct the child to invoke `helpers/epic-deliver-story <storyId>`
-(whose Step 4 defines the child's return shape), (3) remind the child
-of the **non-interactive contract** (no clarifying questions;
-transition to `agent::blocked` and exit if stuck), (4) tell the child to
-suppress per-Story chat relay and instead relay **one line per phase
-transition** (e.g. `Story #<id>: implementing → closing`) — the child's
-authoritative progress lands in the `story-run-progress` snapshot the
-`story-phase.js` CLI upserts, not in a verbatim body dump, (5) remind the
-child that its `story.heartbeat` liveness signal is emitted **off the token
-stream** by the PostToolUse hook (Epic #4476) — a throttled heartbeat lands as
-a free byproduct of every tool call, keyed off the active-Story env
-`story-init.js` exported — so it does **not** run `story-phase.js` per step
-just to heartbeat (a `story-phase.js` call at a genuine Story-level phase
-transition still renders the snapshot and stamps an operator-bearing
-heartbeat), and if it cannot make progress it must transition to
-`agent::blocked` rather than fall silent, and (6) pass the
-**docs digest path** — the `docsDigestPath` field from the
-`epic-deliver-prepare.js` envelope (§ Phase 1 main), which points at
-`temp/epic-<epicId>/docs-digest.md`. Instruct the child to read that
-digest instead of re-reading the full `project.docsContextFiles` set,
-and to pull individual docs files on demand (per
-[`.agents/instructions.md` § 3](../../instructions.md)). When
-`docsDigestPath` is null (the project configured no `docsContextFiles`),
-say so — the child then has no per-Story docs mandate. (7) pass the
-**checklist path** — the `checklistPath` field from that Story's entry in
-the `stories` dispatch hint (Story #4410), which points at
-`temp/epic-<epicId>/checklists/story-<storyId>.md`. Instruct the child to
-read that file (footprint-matched local-lens authoring checklists) and
-self-check its change against those concerns while writing. When
-`checklistPath` is null (the Story's predicted footprint matched no local
-lens), there is nothing to read. The pairing of
-`story.heartbeat` and `agent::blocked` is what lets the § 2d Idle
-Watchdog distinguish a working child from a dead one; a silent child
-with no recent heartbeat and no blocker label is the failure mode the
-watchdog is built to catch.
-
-There is **no per-child JSON return-parsing ceremony** for the parent
-to enforce. GitHub state is the contract: `epic-execute-record-wave.js`
-(§ 2c, mode B) treats each child's raw return text as a best-effort
-hint and reconciles any unparseable, empty, or missing return directly
-from the Story's live labels and comments.
-
-### 2c. Record the Story outcomes
-
-As dispatched Stories return (record them as they land — you need not wait
-for the whole ready set), persist each Story's terminal status via
-`epic-execute-record-wave.js`. There is **no `--wave` flag and no
-`currentWave`** — the recorder splices each Story's status into the
-checkpoint's flat per-Story map and re-renders the rollup:
-
-```bash
-# Mode A — host LLM already parsed each child return.
-node .agents/scripts/epic-execute-record-wave.js \
-  --epic <epicId> --results @<file>|<inline-json>
-
-# Mode B — pipe the raw per-Story sub-agent return texts directly.
-node .agents/scripts/epic-execute-record-wave.js \
-  --epic <epicId> --returns @<file>|<inline-json>
-# `<inline-json>` shape: [{ "storyId": <n>, "returnText": "<raw text>" }]
-```
-
-**Mode B is the default path** — pipe the raw return texts through
-without inspecting them. The CLI reconciles parse failures from GitHub,
-records each Story's terminal status, emits one `story.dispatch.end` per
-recorded Story (closing the ledger pairing), re-renders
-`epic-run-progress`, and prints `{ status, nextAction, renderedBody, ... }`.
-Print `renderedBody` verbatim, then optionally append a short **Notable**
-section (0–5 bullets on newly blocked / failed / slow Stories, friction,
-elapsed-time surprises).
-
-> **Crash recovery.** A child that finished but was never recorded is
-> re-derived from its live label on the next `tick` and never
-> re-dispatched. See
-> [`deliver-epic-reference.md` § Crash recovery (record step)](deliver-epic-reference.md#crash-recovery-record-step)
-> for the manual re-record command.
-
-### 2d. Loop on `nextAction`
-
-After `2c`, re-run `wave-tick.js`. Branch on the new envelope:
-
-- `dispatch` → repeat 2b/2c for the new ready set (the next beat's
-  dependency-satisfied Stories), then re-tick.
-- `observe` → poll the Epic (children may still be in flight, or some
-  are `agent::blocked`). If `blockedStories` is non-empty, post a
-  friction comment, flip Epic to `agent::blocked`, park.
-- `halt` → the run is stuck: no Story is dispatchable, nothing is in
-  flight, yet not every Story is done. `reason` distinguishes the two
-  causes — `dependency-cycle` (the in-scope Stories form a `blocked by`
-  cycle; `cycle` lists the offending Story ids) or
-  `unsatisfiable-dependency` (a Story is gated on a dependency that can
-  never satisfy). `stuckStories` names the Story id(s) that stranded the
-  run. Post a friction comment quoting `reason` + `stuckStories`, flip the
-  Epic to `agent::blocked`, and park for the operator. **Never** treat a
-  `halt` as completion — proceeding to Phase 3 would silently drop the
-  stuck Story.
-- `epic-complete` → **every** in-scope Story is done and nothing is in
-  flight; proceed to Phase 3. (The tick returns `epic-complete` only when
-  the done count equals the in-scope Story count — a stuck Story surfaces
-  as `halt`, not a false `epic-complete`.)
-
-> **Idle Watchdog.** While any Story is in flight, re-tick every 30 minutes
-> with `wave-tick.js --epic <epicId> --check-idle 30` so a silent child
-> (crashed host, lost return) is surfaced as a `wave-stall` and
-> re-dispatched or blocked. The full cadence, staleness test (heartbeat +
-> deterministic branch-commit signal), stall envelope, and the "why 30 not
-> 10" rationale are in
-> [`deliver-epic-reference.md` § 2e. Idle Watchdog](deliver-epic-reference.md#2e-idle-watchdog).
-> Stop the cadence once the tick returns `epic-complete`.
-
----
-
-## Phase 3 — Close-validation
-
-Run lint + test + ratchets against `epic/<epicId>` before opening the PR:
-
-```bash
-node .agents/scripts/evidence-gate.js \
-  --epic-id <epicId> --scope-id <epicId> --gate lint -- npm run lint
-node .agents/scripts/evidence-gate.js \
-  --epic-id <epicId> --scope-id <epicId> --gate test -- npm test
-```
-
-If either gate fails: STOP, fix on a hotfix branch, merge back to the
-Epic branch, restart this phase.
-
-### 3.1 Refresh ratcheted baselines
-
-Inspect the scripts in `.husky/pre-push` (typecheck, lint, maintainability,
-design tokens, dependency audits, bundle-size budgets). Run each against
-the Epic branch; if any drifts, refresh and commit
-`chore(baselines): refresh <name> for Epic #<epicId>`.
-
----
-
-## Phase 4 — Epic-close lens roster
-
-Skip when `--skip-epic-audit`. Otherwise auto-invoke
-[`helpers/epic-audit.md`](epic-audit.md) inline to **resolve** (not walk) the
-slim Epic-close lens roster. The helper runs
-[`epic-audit-prepare.js`](../../scripts/epic-audit-prepare.js) to ask the
-[`selectAudits`](../../scripts/lib/audit-suite/index.js) SDK which lenses fire
-at the `gate3` close gate, **unions in the model-judged risk-routed lenses**
-(Story #3889 — `epic-audit-prepare.js` reads the Epic's `planningRisk`
-envelope off the `epic-plan-state` checkpoint and maps each high-risk axis to
-its lens via `resolveAuditLenses`), then restricts that union to the tiers the
-Epic-close tier owns via
-[`selectEpicCloseLenses`](../../scripts/lib/orchestration/code-review.js): the
-envelope's **`epicCloseLenses`** field is the slim roster of **cumulative +
-global + risk-routed** lenses, with every **local-tier** change-set lens
-excluded (its concern is already verified shift-left at the write-time and
-Story-scope tiers). A high-risk Epic still auto-runs its risk-routed lenses
-even when the change set alone did not select them; a docs-only or
-already-shift-left-covered change set yields an empty roster.
-
-**The lens roster is not walked here.** Story #4412 folded the standalone
-Phase 4 lens walk into the Phase 5 code-review pass so the cumulative Epic diff
-is walked **once**. Phase 4 resolves `epicCloseLenses` (plus `depth`,
-`globalLenses`, `substitutionsPayload`) and hands it to Phase 5; there is no
-separate `audit-results` comment. Remediation of the lens findings happens in
-Phase 5 and is **tier-aware** (Story #4412): the Epic-close tier reads
-`delivery.epicAudit.autoFixSeverity` (default **`high`**) and routes only
-🔴 Critical + 🟠 High lens findings into on-branch remediation while 🟡 Medium
-and 🟢 Suggestion findings graduate (🟡 Medium concerns are already remediated
-shift-left); setting `medium` opts back into routing 🔴/🟠/🟡.
-
-- **Selector reports `degraded: true`** — STOP. Propagate the
-  `reason`/`detail`, post a friction comment, do not fall back to a
-  full-roster audit.
-- **`epicCloseLenses` is empty** (docs-only change set, or every selected lens
-  already covered shift-left, and no risk-routed lens) — there is no lens
-  dimension for Phase 5 to walk; continue to Phase 5, which still runs its
-  review pillars.
-
----
-
-## Phase 5 — Code review (cumulative diff walked once)
-
-Skip when `--skip-code-review`. Otherwise resolve the **risk-derived review
-depth** for this Epic, then auto-invoke
-[`helpers/code-review.md`](code-review.md) inline (read-only audit)
-with the argument envelope `{ scope: 'epic', ticketId: <epicId>, baseRef:
-'main', headRef: 'epic/<epicId>', depth: <reviewDepth> }`, threading the
-Phase 4 `epicCloseLenses` roster through. The pass walks the cumulative
-`main..epic/<epicId>` diff **once**: it executes the Epic-close lens roster as
-review dimensions (helper Step 1b) **and** the review pillars, folding both
-into a single aggregate. Findings persist as the unified `verification-results`
-structured comment on the Epic (the single findings surface — Story #4411
-unified the former `code-review` and `audit-results` contracts, and Story #4412
-folded the lens walk into this pass).
-
-The `depth` is the live epic-scope producer for Story #3876's review-depth
-lever (Story #3937). Resolve it from the Epic's judged risk envelope the same
-best-effort way Phase 4 routes audit lenses — via
-[`resolveReviewDepthForEpic`](../../scripts/lib/orchestration/code-review.js),
-which reads `planningRisk.overallLevel` off the Epic's `epic-plan-state`
-checkpoint and maps it: `high` → `deep`, `low` → `light`, everything else
-(including a missing/unparseable checkpoint, or an Epic that skipped
-`/plan`) → `standard`. The helper threads `depth` into `runCodeReview`,
-which forwards it to every provider's `runReview` input; the LLM-backed
-providers (codex, security-review, ultrareview) render it into the prompt they
-emit so a high-risk Epic gets a deeper adversarial pass and a low-risk one a
-lighter one. Depth is **input-only** — it never changes the findings envelope
-or the posted comment shape.
-
-Remediation in this pass is **tier-aware and split by finding class**
-(Story #4412). The **review-pillar** findings (Pillars 1–4) route off
-`delivery.codeReview.autoFixSeverity` (default **`medium`** — 🔴/🟠/🟡
-on-branch, Mediums batched per lens, 🟢 stays on the comment). The **Epic-close
-lens** findings (Step 1b) route off `delivery.epicAudit.autoFixSeverity`
-(default **`high`** — only 🔴/🟠 on-branch; 🟡 Medium + 🟢 graduate, because
-🟡 Medium concerns were already remediated shift-left). Setting either key to
-its other value re-widens or narrows that class. Remediated findings are
-rendered under the single `verification-results` comment's `## Fixed on-branch`
-section so they never graduate to follow-up issues. The severity gate below is
-**unchanged** — it keys off the surviving (unfixed) findings.
-
-- **Any surviving 🔴 Critical Blocker** — STOP. Relay to the operator.
-- **Only 🟠/🟡/🟢 surviving** — log as non-blocking and continue.
-
----
-
-## Phase 6 — Retro
-
-Skip when `--skip-retro`. Otherwise post the `epic-perf-report` via
-`node .agents/scripts/analyze-execution.js --epic <epicId>` (failure →
-warn and continue; the retro runner falls back). Then invoke the retro
-runner via its CLI wrapper:
-
-```bash
-node .agents/scripts/retro-run.js --epic <epicId>
-```
-
-[`retro-run.js`](../../scripts/retro-run.js) resolves the config/provider,
-constructs a lifecycle bus with a `LedgerWriter` (so the run's
-`retro.start` / `retro.end` boundaries land in
-`temp/epic-<epicId>/lifecycle.ndjson`), and calls `runRetro` — the
-canonical compose-and-post surface at
-[`.agents/scripts/lib/orchestration/retro-runner.js`](../../scripts/lib/orchestration/retro-runner.js).
-Propagate `--full-retro` to bypass the compact-path heuristic.
-
-Retro fires here (before the PR opens) so it stays in the operator's
-local session with full env access (env vars, credentials, MCP). After
-the GitHub upsert succeeds, the retro body is also **mirrored locally** to
-`temp/epic-<epicId>/retro.md` (path resolved via
-[`lib/config/temp-paths.js`](../../scripts/lib/config/temp-paths.js)'s
-`epicRetroMirrorPath`). GitHub remains the source of truth — a
-mirror-write failure only logs a warn and never fails the phase.
-
-**Auto-file routed proposals (Story #4418).** Between gathering the retro
-signals and composing the body, the runner files the retro's **actionable
-routed proposals** (the `framework` / `consumer` friction categories that
-recurred ≥ 2× or were force-flagged by an unresolved `agent::blocked`) as
-GitHub follow-up issues via the graduator pre-parsed-findings seam
-([`retro-proposals-graduator.js`](../../scripts/lib/feedback-loop/retro-proposals-graduator.js)).
-Each filed issue carries `meta::<framework-gap|consumer-improvement>` +
-`friction::<category>` labels — the join key `/plan` Phase 0's
-prior-feedback fetcher reads back — and the rendered retro sections then
-list the **real filed issue numbers** instead of paste-ready `gh issue
-create` command stanzas. Filing is idempotent (a content-hash marker probe
-skips already-filed categories) and respects the graduator per-run filing
-cap. The beat runs behind `delivery.feedbackLoop.retroProposals` (default
-**ON**); set it to `false` to suppress auto-filing and fall back to the
-command stanzas. Filing never fails the phase — a filing error degrades to
-the command stanzas.
-
----
-
-## Phase 6.5 — Post-wave integration gate (Epic #4131, F1/F4)
-
-This phase runs **after** the Phase 2 wave loop reports `epic-complete` and
-**before** the Phase 7 finalize emit opens the PR to `main`. It is the one
-**deliberately-global** gate — its evidence spans the whole product, not just
-the Epic's change set — so it catches the surface each Story shipped correctly
-in isolation yet the assembled product cannot reach (an unnavigable route, a
-broken persona journey).
-
-Skip when `--skip-integration-gate` (log the override; record a manual
-intervention per
-[`deliver-epic-reference.md` § Recording manual interventions](deliver-epic-reference.md#recording-manual-interventions)).
-The gate is otherwise **always evaluated** but a **silent no-op when
-unconfigured** (no `routeGlobs` / `navRegistry` / `journeySuite` in
-`.agentrc.json`).
-
-Sub-steps and hard-failure semantics:
-
-- **6.5a — Whole-product navigability**: run the `navigability` lens in
-  whole-route mode over the `epic/<epicId>` tip. An **orphaned** route (no
-  nav door for any entitled persona) or a dead nav href is a hard failure
-  that **blocks finalize** and names the surface.
-- **6.5b — Consumer journey suite**: run
-  `delivery.quality.navigability.journeySuite` over the tip. A failing
-  persona journey is a hard failure that **blocks finalize** and names the
-  broken journey.
-- **6.5c — `@pending` ≠ green for surface-adding Epics (F4)**: for a
-  **surface-adding** Epic, an AC covered **only** by `@pending` scenarios is
-  treated as unsatisfied and **fails the close gate** instead of passing
-  green. This is **purely additive** and scoped to surface-adding Epics —
-  refactor-only and docs-only Epics are **unaffected** and the existing
-  `satisfied` / `missing` reconciliation is **not de-scoped** for any Epic.
-
-On any hard failure, post a friction structured comment naming the surface
-(route / nav-door identifier only — never route bodies or persona PII per
-`security-baseline.md`), flip the Epic to `agent::blocked`, and **do not**
-open the PR — the gate fails safe and loud. See
-[`deliver-epic-reference.md` § Phase 6.5](deliver-epic-reference.md#phase-65--post-wave-integration-gate-epic-4131-f1f4)
-for the full lens config, the surface-adding-signal derivation, the no-op
-degradation contract, and the fail-safe-and-loud security note.
-
----
-
-## Phase 7 — Finalize (ready the PR / open PR to main)
-
-Before the close-tail emit, sync the Epic branch with `origin/main` so the
-PR opens with the latest base commits already integrated (a stale base
-stalls at branch-protection's `up-to-date branch` rule):
-
-```bash
-git checkout epic/<epicId>
-node .agents/scripts/sync-branch-from-base.js \
-  --branch epic/<epicId> --base main
-git push origin epic/<epicId>
-```
-
-Then fire the close-tail emit:
-
-```bash
-node .agents/scripts/lifecycle-emit.js --epic <epicId> --event epic.close.end
-```
-
-`epic.close.end` drives the bus-owned `Finalizer` chain: acceptance-table
-reconciliation (throws and aborts finalize on a coverage gap, `waived` under
-`acceptance::n-a`), the PR-open/ready step (below), and the
-`epic-handoff` comment naming the PR URL. The chain emits `pr.created` →
-`epic.finalize.end` and **stops** — it never emits `epic.merge.ready` (the
-auto-merge arm is driven later from the Phase 8.5 gated watch path). The
-operator shells nothing beyond the sync and the single emit.
-
-**PR-open/ready is gated by `delivery.ci.earlyPr` (Story #4359).** Resolve
-the flag through the [`getCiDelivery`](../../scripts/lib/config/ci.js)
-accessor (default `true`); do not read `delivery.ci.earlyPr` directly.
-
-- **`earlyPr` on (default)** — the Epic PR already exists as a draft (Phase
-  2 opened it at wave 1). Finalize **locates** the existing PR and flips it
-  ready-for-review via
-  [`markPrReady`](../../scripts/lib/orchestration/finalize/open-or-locate-pr.js)
-  rather than creating a PR. `gh pr ready` on an already-ready PR is a
-  no-op, so a re-run is idempotent.
-- **`earlyPr` off** — no draft was opened at wave 1; finalize opens the PR
-  now via `openOrLocatePr` (no `draft`), exactly as the pre-Story timing.
-
-In both modes the PR title/body contract (`feat: Epic #<epicId>` /
-`Closes #<epicId>`) is identical.
-
-**Drain the bookkeeping outbox (Epic #4476 M5).** When this run was invoked
-with `--yes` (headless), the mechanical bookkeeping — non-urgent
-structured-comment upserts and intermediate `agent::*` label flips — was
-**buffered to a local per-Epic outbox** instead of posting live per
-transition (see the box below). Reconcile it to GitHub once, now, after the PR
-is open:
-
-```bash
-node .agents/scripts/bookkeeping-reconcile.js --epic <epicId>
-```
-
-This drains `temp/epic-<epicId>/bookkeeping-outbox.ndjson` (FIFO,
-idempotent) so GitHub is the source of truth at rest. It exits non-zero and
-**retains** the outbox if any op fails (crash-recovery: a later reconcile
-re-drains the remainder). Attended runs buffered nothing, so this is a no-op.
-
-> **Headless buffering, not silence (§1.H / §1.J).** In `--yes` mode the
-> comment/label CLIs accept `--buffer --epic <epicId>` to enqueue a mutation
-> instead of a live round-trip
-> ([`post-structured-comment.js`](../../scripts/post-structured-comment.js),
-> [`update-ticket-state.js`](../../scripts/update-ticket-state.js)). The
-> `agent::blocked` HITL gate is **never** buffered — a genuine blocker flips
-> the label live and immediately, exactly as before, so the operator can see
-> and resume it. `agent::done` is also never buffered (its cascade runs live).
-> Attended runs pass no `--buffer` and behave byte-for-byte as they always
-> have.
-
-See
-[`deliver-epic-reference.md` § Phase 7 — Finalize](deliver-epic-reference.md#phase-7--finalize-close-tail-listener-chain)
-for the branch-sync outcome table (conflict / fetch-failed recovery) and the
-full three-step listener contract (why finalize must not emit
-`epic.merge.ready`, the merge-lockout lint rule, the no planning-ticket close
-sweep).
-
----
-
-## Phase 8 — Watch-and-iterate until CI is green
-
-The host LLM owns the green-bar loop until the operator merges. Use
-`pr-watch-with-update.js` — the **single CI-watch mechanism** shared with
-the standalone single-Story Step 4 path (Story #4358). It polls the PR's
-required checks to a terminal state and additionally auto-recovers from
-`mergeStateStatus: BEHIND` by calling `gh pr update-branch` once every
-required check is green (branch-protection rules requiring "up to date
-before merging" otherwise park the PR until the operator clicks **Update
-branch** manually):
-
-```bash
-node <agentRoot>/scripts/pr-watch-with-update.js --pr <prNumber> --epic <epicId>
-```
-
-`<agentRoot>` resolves from `project.paths.agentRoot` (default `.agents`).
-Poll cadence and caps come from `delivery.ci.watch.*`
-(`pollIntervalMs`, `maxPolls`, `maxResumes`); pass `--poll-interval-ms`,
-`--max-polls`, `--max-resumes`, or `--max-updates` to override for one
-run. Passing `--epic <epicId>` scopes the red-path failure digest to
-`temp/epic-<epicId>-ci-digest.{json,md}`.
-
-**Three-way exit (slow-vs-failed semantics):**
-
-- **Exit 0** — every required check is green → proceed to Phase 8.5.
-- **Exit 1** — a required check genuinely failed (red). The CLI writes
-  `temp/epic-<epicId>-ci-digest.{json,md}` (failing check, run id,
-  `gh run view --log-failed` tail, coarse classification) and surfaces
-  the fix-loop handoff. Remediate on `epic/<epicId>` and re-run the
-  helper (auto-merge stays armed across retries). If the same failure
-  class recurs, hand the convergence off to a self-paced host loop
-  (`/loop`) that re-runs the failing check and applies the smallest fix
-  until it exits green.
-- **Exit 2** — **still-running** (slow CI, not red): the poll cap fired
-  with checks still pending and the watcher exhausted its
-  `delivery.ci.watch.maxResumes` re-arm budget with nothing red. This is
-  **never** a failure and **never** `timed_out`. Hand the wait off to the
-  host's interval loop rather than blocking the delivery turn: `/loop 5m`
-  polling `gh pr checks` until the checks settle.
-
-> **Triage authority.** How to classify and remediate a red (or repeatedly
-> slow) check — the root-cause-only decision tree for infra/transient and
-> flaky failures (reproduce → check `main` → bisect env vs code → fix in-scope
-> or file a `meta::framework-gap` issue), the never-rerun / never-quarantine
-> prohibitions, and the escalation criteria (three-strikes, the 30-minute
-> wall-clock timebox, and the clearly-environmental fast path) — is defined
-> once in [`.agents/rules/ci-remediation.md`](../../rules/ci-remediation.md).
-> Read it before remediating.
->
-> **Remediation + hard prohibitions.** For the per-check fix table (lint,
-> baseline drift, test, coverage), the three-strikes halt rule, and the
-> never-merge / never-force-push / never-dodge prohibitions, see
-> [`deliver-epic-reference.md` § Phase 8 — Watch-and-iterate remediation](deliver-epic-reference.md#phase-8--watch-and-iterate-remediation).
-
----
-
-## Phase 8.5 — Auto-merge gate
-
-After Phase 8 exits 0, evaluate the auto-merge predicate by emitting
-`epic.automerge.start`. When this `/deliver` run was invoked with `--yes`,
-add `--headless true` so the downstream `MergeWatcher` engages the
-must-land terminal step (Story #4427, § Arguments above); omit the flag
-(or pass `--headless false`) for an attended run:
-
-```bash
-node .agents/scripts/lifecycle-emit.js --epic <epicId> \
-  --event epic.automerge.start --pr-url <prUrl> [--headless true]
-```
-
-**Must-land terminal step (headless only).** `MergeWatcher` polls
-`epic.merge.armed` to confirmation as usual. If its poll budget is
-exhausted, a headless (`--headless true`) run does not exit silently —
-it classifies the block (`classifyMergeBlock`, the shared classifier
-from Story #4426) and applies one bounded retry before giving up:
-
-- `checks-pending-timeout` (required checks still progressing) — extend
-  the watch budget once and keep polling.
-- `api-race-other` (no definitive block signal) — re-arm once by
-  re-emitting `epic.merge.ready` on the bus (never a direct `gh pr merge`
-  call — `AutomergeArmer` remains the sole authorized call site).
-- `branch-protection-human-required`, or both bounded retries already
-  spent — terminal: emit `merge.unlanded` (`scope: "epic"`, carrying the
-  block class) and fall through to the existing single `epic.blocked`
-  emit — one blocked path, never a duplicate `agent::blocked` transition.
-
-Attended runs (no `--headless` flag) keep today's exact behavior: budget
-exhaustion emits `epic.blocked` immediately, with no classification, no
-retry, and no `merge.unlanded`.
-
-`AutomergePredicate` first runs a **live `gh pr checks --required` probe**
-(Story #4361): green required CI is the arming signal, so if any required
-check is red, pending, or the probe is unreadable it emits
-`epic.merge.blocked` immediately — even if the Phase 8 watch was interrupted
-before it observed green (closing the Story #3901 interrupted-watch hole).
-When the probe is green it evaluates the structured-signal verdict under the
-`delivery.ci.autoMerge` policy (default `"trust-ci"`; see
-[`configuration.md`](../../docs/configuration.md)):
-
-- **`trust-ci`** (default) — the ONLY structured conditions that block
-  arming are an unresolved 🔴 critical (red) code-review finding or an
-  `agent::blocked` state (a story-level blocker recorded in run-state, a
-  non-done story, or a missing run-state checkpoint). Manual interventions,
-  🟠 warning-level findings, and a non-clean retro are **recorded for audit**
-  (surfaced on the classification log and the arm-reason) but no longer block.
-- **`strict`** — restores the prior clean-sprint predicate exactly: empty
-  manual-interventions, every story done, no story blocked, `0` 🔴 + `0` 🟠
-  review findings, and the retro's `automerge-verdict` trailer reporting
-  `cleanSprint: true`. Any dirty signal blocks.
-
-On an arming decision the predicate emits `epic.merge.ready`; the downstream
-`AutomergeArmer` (the sole authorized `gh pr merge` call site) fires
-`gh pr merge --auto --squash --delete-branch`. Otherwise the predicate emits
-`epic.merge.blocked` with the disqualifying reasons and exits without merging
-— the operator merges manually.
-
-**Blocked-path output (operator merges the button).** When arming is
-declined, `epic.merge.armed` never fires inside this run, so Phase 9 does not
-reap automatically. Surface the exact one-liner the operator runs **after**
-they merge the PR by hand so local refs are reaped and `main` is
-fast-forwarded (the idempotent-resume path below runs this automatically on
-the next `/deliver <epicId>`):
-
-```bash
-node .agents/scripts/lifecycle-emit.js --epic <epicId> \
-  --event epic.merge.armed --pr-url <prUrl>
-```
-
-Close the phase wrapper by emitting `epic.automerge.end` (records the arm
-outcome on the ledger; `merged: true` once GitHub completes the squash,
-`merged: false` with a reason otherwise):
-
-```bash
-node .agents/scripts/lifecycle-emit.js --epic <epicId> \
-  --event epic.automerge.end --pr-url <prUrl> --merged <true|false>
-```
-
-> **Predicate wiring + manual-intervention recording.** For the full
-> predicate contract (the trailer read, the `delivery.ci.autoMerge` policy
-> split, and the Story #4361 live `gh pr checks --required` probe that
-> replaced the former CI-freshness skip) and the
-> `epic-deliver-note-intervention.js` command + its trigger list, see
-> [`deliver-epic-reference.md` § Phase 8.5 — Auto-merge predicate detail](deliver-epic-reference.md#phase-85--auto-merge-predicate-detail).
-
----
-
-## Phase 9 — Local branch cleanup
-
-Phase 9 runs **automatically** inside the lifecycle bus once auto-merge
-arms: the `BranchCleaner` listener subscribes to `epic.cleanup.start`
-and reaps local refs (the `epic/<id>` branch, every `story-<id>` in the
-checkpoint, attached worktrees, and stale tracking refs) before `Cleaner`
-archives the `temp/epic-<id>/` tree. No operator step is required on the
-auto-merge path.
-
-For out-of-band cleanup re-entry (resume after a crash, or operator
-override), fire `epic.merge.armed`:
-
-```bash
-node .agents/scripts/lifecycle-emit.js --epic <epicId> \
-  --event epic.merge.armed --pr-url <prUrl>
-```
-
-> **Reap order + operator-merges fallback.** For the full in-process reap
-> order, the per-branch classification log, and the manual reap sequence
-> when Phase 8.5 fell back to the operator-merges-button path (auto-merge
-> declined, `epic.merge.armed` never fired), see
-> [`deliver-epic-reference.md` § Phase 9 — Local branch cleanup detail](deliver-epic-reference.md#phase-9--local-branch-cleanup-detail).
-
----
-
-## Idempotence and resume
-
-Re-runs pick up at the next undispatched wave (in-flight Stories finish
-via `helpers/epic-deliver-story`'s own checkpointing). The PR from Phase 7 is
-updated in place on subsequent runs. The authoritative live view is
-the `epic-run-progress` structured comment.
-
-**Resume auto-arm for a merged-but-uncleaned Epic.** When `/deliver` resumes
-against an Epic whose PR already merged (operator merged the button in a prior
-session) but whose local `epic/<id>` / `story-<id>` refs still linger, the
-resume path detects the merged-but-uncleaned state
-(`detectMergedUncleanedEpic` in
-[`epic-cleanup.js`](../../scripts/lib/orchestration/epic-cleanup.js)) and fires
-`epic.merge.armed` automatically so Phase 9 reaps — no manual command. The
-detection is idempotent: an already-reaped Epic (no local refs) is a clean
-no-op, and an unmerged Epic never arms. It resolves the merged PR's URL for the
-required `epic.merge.armed` payload and fails closed (does **not** arm) on any
-indeterminate `gh` probe. The one-liner under Phase 8.5 / Phase 9 is the manual
-equivalent for the case where the operator does not re-run `/deliver`.
-
----
-
-## Constraints
-
-- **Never** merge `epic/<epicId>` to `main` outside Phase 8.5.
-- **Never** dispatch more than the global `concurrencyCap` allows;
-  concurrency lives inside the ready-set fan-out.
-- **Never** flip Story-level labels from this skill; **never** invoke
-  `helpers/epic-deliver-story` yourself (children run it via Agent fan-out,
-  even for single-Story waves); **never** spawn a subprocess for dispatch.
-- **Always** checkpoint via `epic-deliver-prepare.js` /
-  `epic-execute-record-wave.js`; never write run state elsewhere.
-- **Always** post a friction structured comment before a non-`complete`
-  outcome.
-- **Always** auto-invoke the epic-audit, code-review, and retro helpers
-  (Phases 4–6) when their artefacts aren't already present.
-- **Always** run the Phase 6.5 integration gate after the wave loop
-  reports `epic-complete` and before Phase 7 finalize (unless
-  `--skip-integration-gate`); **never** open the PR while the gate
-  reports a hard failure (orphaned surface, dead nav href, broken
-  journey, or a surface-adding Epic with only `@pending` AC coverage).
-- **Always** drive Phase 8 to green CI before returning control — the
-  host LLM owns the loop until the PR is mergeable or the Epic is
-  parked at `agent::blocked`.