peaks-loop 4.0.46 → 4.0.48

package/skills/bee/peaks-rd/SKILL.md

@@ -9,7 +9,7 @@ description: |
 visibility: internal
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 
@@ -172,22 +172,24 @@ If any gate fails, return to development for fixes or hand off as blocked. Do no
 
 **v2.12.0 collapse (Group A — Tier 1+2+3):** the previous 5-way fan-out (slice 004 4-way + slice 5/6 `karpathy-reviewer` addition) totalled **5 sub-agents**. The `security-reviewer` and `perf-baseline-reviewer` slots moved out of the RD 3-way fan-out into two new standalone audit skills:
 
-- `peaks-security-audit` — CLI: `peaks security-audit run`. Writes `audit/security.md` (under `.peaks/_runtime/<sessionId>/audit/security.md`). Required RD-side prereq `AUDIT_SECURITY`.
-- `peaks-perf-audit` — CLI: `peaks perf-audit run`. Writes `audit/perf.md` (under `.peaks/_runtime/<sessionId>/audit/perf.md`). Required RD-side prereq `AUDIT_PERF`.
+- `peaks-security-audit` — CLI: `peaks security-audit run`. Writes `audit/security-<rid>.md` (under `.peaks/_runtime/<sessionId>/audit/security-<rid>.md`). Required RD-side prereq `AUDIT_SECURITY`.
+- `peaks-perf-audit` — CLI: `peaks perf-audit run`. Writes `audit/perf-<rid>.md` (under `.peaks/_runtime/<sessionId>/audit/perf-<rid>.md`). Required RD-side prereq `AUDIT_PERF`.
+
+  **The rid is part of the filename** (slice `2026-09-14-audit-artifact-rid-scoping`). Two slices run in the same session share `.peaks/_runtime/<sessionId>/audit/`, so a ridless filename makes the second slice's write silently destroy the first slice's audit — which is what happened to `audit/perf.md` on 2026-09-13, irrecoverably. The bare `audit/security.md` / `audit/perf.md` (and the older `rd/security-review.md` / `rd/perf-baseline.md`) are still **read** as back-compat fallbacks, but never write there.
 
 Both audit skills consume the immutable peaks-prd handoff (`prd/handoff.md`) and the project-scoped audit templates under `.peaks/project-scan/{security-template, perf-template, audit-output-schema}.md`. The handoff presence is enforced by the `AUDIT_REQUIRES_HANDOFF` prereq. The 1-minor-release back-compat window (v2.12.0) keeps the old `rd/security-review.md` and `rd/perf-baseline.md` paths readable via `mustContainAny` — see `references/rd-fanout-contracts.md` §"Deprecated reviewer back-compat".
 
 **Current 3-way fan-out** (always runs for feature / refactor / bugfix; no fan-out for config / docs / chore):
 
-1. `code-reviewer` — writes `rd/code-review.md` (Gate B3).
+1. `code-reviewer` — writes `rd/code-review-<rid>.md` (Gate B3). **Not** `rd/code-review.md`: the ridless name is shared by every slice in the session, so slice 2 would overwrite slice 1's review (it did on 2026-09-13). The ridless path is still read as a back-compat fallback.
 2. `qa-test-cases-writer` — writes `qa/test-cases/<rid>.md` (Gate C2).
-3. `karpathy-reviewer` — writes `rd/karpathy-review.md` (the **hard Karpathy-Gate**, KARPATHY_REVIEW prereq).
+3. `karpathy-reviewer` — writes `rd/karpathy-review-<rid>.md` (the **hard Karpathy-Gate**, KARPATHY_REVIEW prereq). Same rule: the ridless `rd/karpathy-review.md` is read but never written.
 
 Full dispatch contract (when-to-fan-out rules, dispatch template, prereq gates) lives in **`references/parallel-review-fanout.md`**. Read that file before issuing any 3-way fan-out.
 
 ## Reviewer fan-out detached mode (Phase C, slice 2026-08-10)
 
-Reviewer fan-out may run in detached mode for any of the 3-way fan-out roles (`code-reviewer`, `qa-test-cases-writer`, `karpathy-reviewer`). Detached mode spawns a real OS process via `peaks sub-agent dispatch --mode detached --vendor <vendor>`, isolated from the orchestrator session. Use detached mode when the reviewer's expected runtime exceeds 60s OR processes ≥ 20 source files. Pass `--mode detached` explicitly per reviewer role; the default remains `in-process` (backward compat). The karpathy-reviewer's G8 infinite-context marker (auto-compact at 0.85/0.95, unlimited spend) is honored automatically when `--mode detached` is used. See `peaks-code/references/sub-agent-dispatch.md` §"Detached Mode" for the full contract.
+Reviewer fan-out may run in detached mode for any of the 3-way fan-out roles (`code-reviewer`, `qa-test-cases-writer`, `karpathy-reviewer`). Detached mode spawns a real OS process via `peaks sub-agent dispatch --mode detached --vendor <vendor>`, isolated from the orchestrator session. Use detached mode when the reviewer's expected runtime exceeds 60s OR processes ≥ 20 source files. Pass `--mode detached` explicitly per reviewer role; the default remains `in-process` (backward compat). The karpathy-reviewer's `<peaks-auto-compact>` marker is honored automatically when `--mode detached` is used. Note what that marker does and does not do (corrected 2026-09-13, rid `2026-09-13-defects-e`): it asks the child to persist durable state to `.peaks/_runtime/<sid>/detached/<rid>/compact/<n>.json` before its own harness compacts the session — the child cannot compact itself, and nothing is spliced back into its prompt. See `peaks-code/references/sub-agent-dispatch.md` §"Detached Mode" for the full contract.
 
 ## Refactor hard gates
 
@@ -269,11 +271,11 @@ Do not bypass PRD/QA artifacts. Do not install hooks, agents, MCP, or settings.
 
 Do not bypass the parallel review fan-out when the slice has a code-review / qa-test-cases / karpathy-review surface — see `## Parallel review fan-out` above. The three review activities are fan-out, not sequential; sequential re-implementation of the same logic by the main RD loop defeats the wall-clock benefit and is treated as a red-line violation.
 
-**Security / perf audit boundary (v2.12.0):** security and perf audit run as standalone audit skills (`peaks-security-audit`, `peaks-perf-audit`) whose outputs land at `audit/security.md` / `audit/perf.md`. RD does **not** dispatch `security-reviewer` or `perf-baseline-reviewer` sub-agents. See `references/rd-fanout-contracts.md` §"Deprecated reviewer back-compat".
+**Security / perf audit boundary (v2.12.0):** security and perf audit run as standalone audit skills (`peaks-security-audit`, `peaks-perf-audit`) whose outputs land at `audit/security-<rid>.md` / `audit/perf-<rid>.md`. RD does **not** dispatch `security-reviewer` or `perf-baseline-reviewer` sub-agents. See `references/rd-fanout-contracts.md` §"Deprecated reviewer back-compat".
 
 ## Karpathy cost self-review (slice 2026-07-30-karpathy-cost-self-review)
 
-The `karpathy-reviewer` sub-agent now reports its own runtime cost in the JSON envelope at `rd/karpathy-review.md` (fields `evaluationCost` + `costRatio`; see `agents/karpathy-reviewer.md` §4). The **orchestrator-side** command `peaks job karpathy-cost-check --review-file <path>` reads that envelope and decides whether to downgrade a `'block'` gateAction to `'warn'` when `costRatio > 10`. **24h-mode is the override** — the check is entirely skipped when `peaks session 24h-mode state` reports `24H_ACTIVE`.
+The `karpathy-reviewer` sub-agent now reports its own runtime cost in the JSON envelope at `rd/karpathy-review-<rid>.md` (fields `evaluationCost` + `costRatio`; see `agents/karpathy-reviewer.md` §4). The **orchestrator-side** command `peaks job karpathy-cost-check --review-file <path>` reads that envelope and decides whether to downgrade a `'block'` gateAction to `'warn'` when `costRatio > 10`. **24h-mode is the override** — the check is entirely skipped when `peaks session 24h-mode state` reports `24H_ACTIVE`.
 
 The main RD loop MUST call `peaks job karpathy-cost-check` after every `peaks request transition --state qa-handoff` and before the next-slice work begins. If the decision kind is `downgraded`, the LLM MUST honor the downgraded `warn` and proceed to the next slice; the `'block'` was an artifact of the reviewer's own cost, not the slice's quality. If the decision kind is `reported` (costRatio > 50, gate not `'block'`), the LLM MAY continue; the sediment will be appended by `peaks memory extract` at handoff time. The full design lives in `.peaks/memory/2026-07-30-karpathy-evaluation-cost-self-review-design.md` (sediment locked 2026-07-30).
 

package/skills/bee/peaks-rd/references/artifact-per-request.md

@@ -99,8 +99,8 @@ For each slice in this request:
 |---|---|---|---|
 | `.peaks/_runtime/<sessionId>/prd/handoff.md` | per-slice — immutable peaks-prd source of truth (v2.11.0+) | RD, QA, all sub-agents | Goals, non-goals, acceptance criteria, architecture, slice graph, mock strategy, cross-cutting decisions. sha256-hashed in frontmatter; sub-agents verify the hash before reading. |
 | `.peaks/_runtime/<sessionId>/rd/requests/<rid>.md` | per-slice — one request, one planning artifact | QA, SC, the lint gate | Red-line scope, in-scope / out-of-scope, unit-test requirements, **Implementation evidence** (file list, `pnpm test` output, git diff excerpts), MCP usage, handoff, status. **This is the file the lint gate checks for placeholders.** |
-| `.peaks/_runtime/<sessionId>/rd/code-review.md` | per-session — the engineering review | QA, the human reviewer | Code review findings + fixes. |
-| `.peaks/_runtime/<sessionId>/rd/security-review.md` | per-session — the security review | QA | Security review findings + fixes. |
+| `.peaks/_runtime/<sessionId>/rd/code-review-<rid>.md` | per-slice — the engineering review | QA, the human reviewer | Code review findings + fixes. |
+| `.peaks/_runtime/<sessionId>/audit/security-<rid>.md` | per-slice — the independent security audit | QA | Security audit findings + verdict. |
 
 > **v2.11.0 change (Group A):** `rd/tech-doc.md` is removed. The per-slice source of truth moves to the immutable peaks-prd handoff (`prd/handoff.md`); the per-slice planning record is `rd/requests/<rid>.md`. The "per-session" content category is no longer RD's responsibility — it lives upstream in the PRD handoff.
 

package/skills/bee/peaks-rd/references/parallel-review-fanout.md

@@ -4,8 +4,10 @@
 
 **v2.12.0 collapse (Group A — Tier 1+2+3):** the previous 4-way fan-out (slice 004) plus the appended `karpathy-reviewer` (slice 5/6) totalled **5 sub-agents**. The `security-reviewer` and `perf-baseline-reviewer` slots moved out of the RD fan-out into two new standalone audit skills:
 
-- `peaks-security-audit` — CLI: `peaks security-audit run`. Writes `.peaks/_runtime/<sessionId>/audit/security.md`. Required RD-side prereq `AUDIT_SECURITY`.
-- `peaks-perf-audit` — CLI: `peaks perf-audit run`. Writes `.peaks/_runtime/<sessionId>/audit/perf.md`. Required RD-side prereq `AUDIT_PERF`.
+- `peaks-security-audit` — CLI: `peaks security-audit run`. Writes `.peaks/_runtime/<sessionId>/audit/security-<rid>.md`. Required RD-side prereq `AUDIT_SECURITY`.
+- `peaks-perf-audit` — CLI: `peaks perf-audit run`. Writes `.peaks/_runtime/<sessionId>/audit/perf-<rid>.md`. Required RD-side prereq `AUDIT_PERF`.
+
+The rid is part of the filename (slice `2026-09-14-audit-artifact-rid-scoping`): every slice in a session shares `.peaks/_runtime/<sessionId>/audit/`, so a ridless name means the second slice's audit silently replaces the first slice's. The ridless locations are still read as fallbacks.
 
 Both audit skills consume the immutable peaks-prd handoff (`prd/handoff.md`) and the project-scoped audit templates under `.peaks/project-scan/{security-template, perf-template, audit-output-schema}.md`. The handoff presence is enforced by the `AUDIT_REQUIRES_HANDOFF` prereq. The 1-minor-release back-compat window (`v2.12.0`) keeps the old `rd/{security-review,perf-baseline}.md` paths readable via `mustContainAny` — see `tests/unit/rd/deprecated-reviewer-back-compat.test.ts` (8 cases) and `tests/unit/artifact-prerequisites-typed.test.ts`.
 
@@ -44,14 +46,14 @@ Note: sub-agent 1 (code-reviewer) and sub-agent 3 (karpathy-reviewer) write to `
 - Read the git diff for this slice (`git diff main...HEAD` or equivalent).
 - Read `.peaks/_runtime/<sessionId>/prd/handoff.md` for slice intent (v2.11.0: the immutable peaks-prd handoff replaces `rd/tech-doc.md`). Verify the handoff hash matches the dispatched value before proceeding.
 - Inspect for: correctness, type safety, error handling, mutation patterns, file-size, naming, dead code, regressions, contract drift.
-- Output: `.peaks/_runtime/<sessionId>/rd/code-review.md` with sections: Summary, Findings, Required Fixes, Recommended, Verdict.
+- Output: `.peaks/_runtime/<sessionId>/rd/code-review-<rid>.md` with sections: Summary, Findings, Required Fixes, Recommended, Verdict. (Rid in the filename — a ridless `rd/code-review.md` is shared by every slice in the session and gets overwritten by the next one; it is still read as a fallback.)
 - Required for Gate B3.
 - **v2.11.0 Tier 7 (Group D) + 2026-09-09-ecc-dynamic:** the code-reviewer dispatch goes through the **ECC bridge** (`src/services/code-review/ecc-bridge.ts`). Fallback order is **native plugin → cache-backed generic agent → inline**:
   1. **Native plugin** (`detectEcc` state `ready`): invoke the Agent tool with `subagent_type: "ecc:code-reviewer"` (plugin `ecc` + agent `code-reviewer`; see `DEFAULT_NATIVE_ECC_AGENT_ID`); it returns the structured envelope `{ passed, violations[], gateAction }`.
   2. **Cache-backed generic agent** (`detectEcc` state `ready-via-cache` — plugin or its review agent absent, but a materialized ECC agent exists under `~/.peaks/agents/ecc/`): resolve the agent's REAL name first — `resolveMaterializedAgentName(['code-reviewer', 'code-review'])` from `peaks-loop-mut` (upstream ships `code-reviewer.md`; a hardcoded `code-review.md` never resolves). Read `~/.peaks/agents/ecc/<resolved>.md`, pass the name through as `detectEcc({ ..., cacheAgentName: resolved })`, build the prompt with `buildCacheBackedEccPrompt({ rid, instructions, diff })` (agent body + diff + `ECC_OUTPUT_CONTRACT`), dispatch a **generic** sub-agent, and validate its reply with `isEccEnvelope`. No ECC plugin required. If the cache is empty, run `peaks ecc install` first (dynamic acquisition); if that fails offline, degrade to inline.
   3. **Inline** (`detectEcc` states `plugin-missing` / `agent-missing` / `dispatch-failed` / `envelope-malformed`): fall back to inline review — TXT note `code-review-ecc-degraded-to-inline`.
 
-  In all dispatchable cases the envelope is rendered by the SAME bridge adapter (`adaptEccEnvelopeToRdCodeReview`) into the canonical `rd/code-review.md` markdown shape that Gate B3 reads (`mustContain: ['## Findings', 'CRITICAL']`). The materialized copy lives under `~/.peaks/agents/ecc/` — peaks-loop NEVER writes into `~/.claude/`.
+  In all dispatchable cases the envelope is rendered by the SAME bridge adapter (`adaptEccEnvelopeToRdCodeReview`) into the canonical `rd/code-review-<rid>.md` markdown shape that Gate B3 reads (`mustContain: ['## Findings', 'CRITICAL']`). The materialized copy lives under `~/.peaks/agents/ecc/` — peaks-loop NEVER writes into `~/.claude/`.
 
 **Sub-agent 2 — qa-test-cases-writer (always runs for feature / refactor / bugfix):**
 - Read the git diff and the PRD acceptance criteria.
@@ -63,7 +65,7 @@ Note: sub-agent 1 (code-reviewer) and sub-agent 3 (karpathy-reviewer) write to `
 **Sub-agent 3 — karpathy-reviewer (always runs for feature / refactor / bugfix — the hard gate):**
 - Inspect the diff + handoff against the 4 Karpathy-guidelines.
 - Read `.peaks/_runtime/<sessionId>/prd/handoff.md` (v2.11.0: architecture summary — the immutable peaks-prd handoff replaces `rd/tech-doc.md`).
-- Output: `.peaks/_runtime/<sessionId>/rd/karpathy-review.md` containing a `## Karpathy-Gate` header and the 4 guideline section markers (Think Before Coding / Simplicity First / Surgical Changes / Goal-Driven Execution).
+- Output: `.peaks/_runtime/<sessionId>/rd/karpathy-review-<rid>.md` containing a `## Karpathy-Gate` header and the 4 guideline section markers (Think Before Coding / Simplicity First / Surgical Changes / Goal-Driven Execution). (Rid in the filename, same reason as the code review; the ridless `rd/karpathy-review.md` is still read as a fallback.)
 - Required for the `KARPATHY_REVIEW` prereq. The transition CLI gate reads those markers and refuses `rd:qa-handoff` when the file is missing or the markers are absent.
 - See `references/rd-fanout-contracts.md` §"karpathy-reviewer contract" for the JSON envelope shape + file format.
 

package/skills/bee/peaks-rd/references/rd-fanout-contracts.md

@@ -11,9 +11,9 @@ end of implementation, RD fires 3 sub-agents in parallel via
 > skills consumed at pre-RD / pre-QA time:
 >
 > - `peaks-security-audit` — CLI: `peaks security-audit run`. Output:
->   `audit/security.md`. Required RD-side prereq: `AUDIT_SECURITY`.
+>   `audit/security-<rid>.md`. Required RD-side prereq: `AUDIT_SECURITY`.
 > - `peaks-perf-audit` — CLI: `peaks perf-audit run`. Output:
->   `audit/perf.md`. Required RD-side prereq: `AUDIT_PERF`.
+>   `audit/perf-<rid>.md`. Required RD-side prereq: `AUDIT_PERF`.
 >
 > Both audit skills consume the immutable peaks-prd handoff
 > (`prd/handoff.md`) and the project-scoped audit templates under
@@ -29,7 +29,7 @@ end of implementation, RD fires 3 sub-agents in parallel via
 > **Karpathy pointer (Slice 1/6):** Each of the 3 sub-agents below operates under the 4 Karpathy guidelines. The canonical reference is `andrej-karpathy-skills:karpathy-guidelines` (full text) and `peaks-rd/SKILL.md` §"Karpathy enforcement". The dispatch primitive also injects the verbatim context block from `rd-sub-agent-dispatch.md` §"Karpathy-guidelines context" into every sub-agent prompt. Sub-agents MUST NOT silently drop the block.
 
 - **Sub-agent 1 — code-reviewer** runs `code-review` against the diff and
-  writes `rd/code-review.md`. **v2.11.0 Tier 7 (Group D) + 2026-09-09-ecc-dynamic:**
+  writes `rd/code-review-<rid>.md`. **v2.11.0 Tier 7 (Group D) + 2026-09-09-ecc-dynamic:**
   the dispatch goes through the **ECC bridge** (`src/services/code-review/ecc-bridge.ts`)
   with fallback order **native plugin → cache-backed generic agent → inline**:
   - `detectEcc` state `ready` → `Agent({ subagent_type: 'ecc:code-reviewer', ... })`
@@ -58,22 +58,22 @@ end of implementation, RD fires 3 sub-agents in parallel via
   writer's only write target is `qa/test-cases/<rid>.md`.
 - **Sub-agent 3 — karpathy-reviewer** (Slice 5/6 — hard gate) inspects
   the diff + handoff against the 4 Karpathy-guidelines and writes
-  `rd/karpathy-review.md` (v2.11.0: the immutable peaks-prd handoff
+  `rd/karpathy-review-<rid>.md` (v2.11.0: the immutable peaks-prd handoff
   at `prd/handoff.md` replaces `rd/tech-doc.md`). The file MUST
   contain a `## Karpathy-Gate` header and at least one of the 4
   guideline section markers; the transition CLI gate reads those
   markers and refuses `rd:qa-handoff` when the file is missing or
   the markers are absent. The sub-agent returns a JSON envelope
   `{ passed, violations, gateAction }` (see contract below). Do NOT
-  modify code; the writer's only write target is `rd/karpathy-review.md`.
+  modify code; the writer's only write target is `rd/karpathy-review-<rid>.md`.
 
 > **Removed from v2.12.0 fan-out (back-compat window only):**
 > - ~~Sub-agent — security-reviewer~~ — moved to standalone
->   `peaks-security-audit` skill; output `audit/security.md`. The legacy
+>   `peaks-security-audit` skill; output `audit/security-<rid>.md`. The legacy
 >   path `.peaks/_runtime/<sessionId>/rd/security-review.md` remains
 >   readable via `mustContainAny` for the v2.12.0 1-minor-release window.
 > - ~~Sub-agent — perf-baseline-reviewer~~ — moved to standalone
->   `peaks-perf-audit` skill; output `audit/perf.md`. The legacy path
+>   `peaks-perf-audit` skill; output `audit/perf-<rid>.md`. The legacy path
 >   `.peaks/_runtime/<sessionId>/rd/perf-baseline.md` remains readable
 >   via `mustContainAny` for the v2.12.0 1-minor-release window.
 >
@@ -103,7 +103,7 @@ on the produced artifacts, and only then attempts
 `peaks request transition --state qa-handoff`. The aggregation step
 runs 3 ls checks: Gate B3 (code-review file), Gate C2 (qa-test-cases
 pre-draft, the 2nd sub-agent's deliverable), and the KARPATHY_REVIEW
-prereq (the 3rd sub-agent's `rd/karpathy-review.md`). The audit
+prereq (the 3rd sub-agent's `rd/karpathy-review-<rid>.md`). The audit
 prereqs (`AUDIT_SECURITY` + `AUDIT_PERF` + `AUDIT_REQUIRES_HANDOFF`)
 are NOT fan-out outputs — they are produced by the standalone audit
 skills and consumed by the CLI gate. A failure in any of the 3
@@ -140,7 +140,7 @@ karpathy §1 Think Before Coding + §3 Surgical Changes.
   any violation is detected, `gateAction` is `warn`. When clean,
   `gateAction` is `pass`.
 
-**File write**: the sub-agent writes ONLY `rd/karpathy-review.md`,
+**File write**: the sub-agent writes ONLY `rd/karpathy-review-<rid>.md`,
 formatted as:
 
 ```md
@@ -208,9 +208,9 @@ failing the gate.
 
 | Request type | Required RD evidence (under `.peaks/_runtime/<sessionId>/`) |
 |---|---|
-| feature / refactor | `prd/handoff.md` (immutable) + `audit/security.md` (peaks-security-audit) + `audit/perf.md` (peaks-perf-audit) + `rd/code-review.md` + `rd/karpathy-review.md` + `qa/test-cases/<rid>.md` |
-| bugfix | `prd/handoff.md` (immutable) + `audit/security.md` (peaks-security-audit) + `audit/perf.md` (peaks-perf-audit, perf-shaped only) + `rd/code-review.md` + `rd/karpathy-review.md` + `qa/test-cases/<rid>.md` |
-| config | `audit/security.md` (peaks-security-audit) |
+| feature / refactor | `prd/handoff.md` (immutable) + `audit/security-<rid>.md` (peaks-security-audit) + `audit/perf-<rid>.md` (peaks-perf-audit) + `rd/code-review-<rid>.md` + `rd/karpathy-review-<rid>.md` + `qa/test-cases/<rid>.md` |
+| bugfix | `prd/handoff.md` (immutable) + `audit/security-<rid>.md` (peaks-security-audit) + `audit/perf-<rid>.md` (peaks-perf-audit, perf-shaped only) + `rd/code-review-<rid>.md` + `rd/karpathy-review-<rid>.md` + `qa/test-cases/<rid>.md` |
+| config | `rd/security-review.md` (genuinely ridless — `SECURITY_REVIEW.relativePath` carries no `<rid>`) |
 | docs / chore | (no extra evidence required) |
 
 Always required (in addition to the type-specific row):
@@ -225,6 +225,6 @@ DO NOT attempt the qa-handoff transition; CLI will reject with
 >
 > **v2.12.0 change (Group A — Tier 4+5):** `rd/security-review.md` and
 > `rd/perf-baseline.md` are removed from the required-evidence matrix;
-> `audit/security.md` (peaks-security-audit) + `audit/perf.md`
+> `audit/security-<rid>.md` (peaks-security-audit) + `audit/perf-<rid>.md`
 > (peaks-perf-audit) replace them. The `AUDIT_REQUIRES_HANDOFF` prereq
 > enforces the immutable handoff consumption by the audit skills.

package/skills/bee/peaks-rd/references/rd-runbook.md

@@ -131,9 +131,13 @@ peaks codegraph affected --project <repo> <changed-files...> --json
 #     writes). Falls back to empty diff if git is unavailable (degraded mode).
 
 # 1.z Karpathy scan (Slice 5/6 + Slice 6/6 — `peaks scan karpathy` + karpathy-reviewer sub-agent)
-#     After the slice is implemented, scan `rd/karpathy-review.md` for the
+#     After the slice is implemented, scan `rd/karpathy-review-<rid>.md` for the
 #     4 Karpathy guidelines (Think / Simplicity / Surgical / Goal) and verify
 #     the hard Karpathy-Gate (KARPATHY_REVIEW prereq in artifact-prerequisites.ts).
+#     NOTE: `peaks scan karpathy` itself has no rid and probes the ridless
+#     back-compat name `rd/karpathy-review.md`; it reports `block` for a slice
+#     that wrote only the rid-scoped name. `peaks request transition` is the
+#     authoritative gate.
 #     The structural scanner covers regex / file-presence checks; the semantic
 #     review is owned by the karpathy-reviewer sub-agent.
 #
@@ -142,7 +146,7 @@ peaks codegraph affected --project <repo> <changed-files...> --json
 #       peaks scan karpathy --project <path> --format json         # machine-readable
 #       peaks scan karpathy --project <path> --scope all           # full audit (gateAction: block if missing)
 #     Required output before `peaks request transition --state qa-handoff`:
-#       - `rd/karpathy-review.md` exists with `## Karpathy-Gate` header
+#       - `rd/karpathy-review-<rid>.md` exists with `## Karpathy-Gate` header
 #       - 4 title-case section headers present (Think Before Coding / Simplicity First / Surgical Changes / Goal-Driven Execution)
 #       - `gateAction: pass` (or `warn` with documented justification)
 #
@@ -153,7 +157,7 @@ peaks codegraph affected --project <repo> <changed-files...> --json
 #     which is the project-internal draft at
 #     `skills/peaks-rd/references/karpathy-reviewer-prompt.md` plus a
 #     `rd/karpathy-reviewer-agent-handoff.md` install guide.
-#     Hard gate: missing karpathy-reviewer sub-agent OR missing rd/karpathy-review.md
+#     Hard gate: missing karpathy-reviewer sub-agent OR missing rd/karpathy-review-<rid>.md
 #     → `peaks request transition --state qa-handoff` returns `code: PREREQUISITES_MISSING`.
 #     Escape hatch (assisted mode): `--allow-incomplete --confirm`.
 
@@ -163,8 +167,8 @@ peaks codegraph affected --project <repo> <changed-files...> --json
 
 # 7. AFTER implementation, BEFORE QA handoff — RUN THESE GATES:
 #    Peaks-Loop Gate B2: unit tests exist and pass for the changed surface → npx vitest run --changed (or project equivalent; the changed-only mode is the peaks slice check default as of run 017; use --run-tests for the full suite, or invoke /peaks-code-test to run the full suite standalone)
-#    Peaks-Loop Gate B3: code review evidence → .peaks/_runtime/<sessionId>/rd/code-review.md
-#    Peaks-Loop Gate B4: security review evidence → .peaks/_runtime/<sessionId>/rd/security-review.md
+#    Peaks-Loop Gate B3: code review evidence → .peaks/_runtime/<sessionId>/rd/code-review-<rid>.md
+#    Peaks-Loop Gate B4: security review evidence → .peaks/_runtime/<sessionId>/audit/security-<rid>.md
 #    Peaks-Loop Gate B5 (NEW): RD artifact body has no unfilled placeholders.
 peaks request lint <rid> --role rd --project <repo> --session-id <session-id> --json
 #    Peaks-Loop Gate B6 (NEW): declared --type still matches the actual diff after implementation.

package/skills/bee/peaks-rd/references/rd-transition-gates.md

@@ -10,9 +10,11 @@ You cannot declare a phase complete from memory. Each gate below is a `ls` or `g
 >
 > | Type | rd:implemented requires | rd:qa-handoff also requires |
 > |---|---|---|
-> | feature / refactor | `rd/tech-doc.md` | `rd/code-review.md` + `rd/security-review.md` + `rd/perf-baseline.md` (filled Results table, or `N/A — no perf surface` in Notes) + **`qa/test-cases/<rid>.md`** (added in slice 004; pre-drafted by the 4th sub-agent in the parallel fan-out) |
-> | bugfix | `rd/bug-analysis.md` (lighter than tech-doc; root cause + fix + regression test plan) | `rd/code-review.md` + `rd/security-review.md` + **`qa/test-cases/<rid>.md`**; `rd/perf-baseline.md` only when the bug is performance-shaped (matches the L449-452 "When this applies" criteria) |
-> | config | (none) | `rd/security-review.md` only |
+> | feature / refactor | `rd/tech-doc.md` | `rd/code-review-<rid>.md` + `audit/security-<rid>.md` + `audit/perf-<rid>.md` (filled Results table, or `N/A — no perf surface` in Notes) + **`qa/test-cases/<rid>.md`** (added in slice 004; pre-drafted by the 4th sub-agent in the parallel fan-out) |
+> | bugfix | `rd/bug-analysis.md` (lighter than tech-doc; root cause + fix + regression test plan) | `rd/code-review-<rid>.md` + `audit/security-<rid>.md` + **`qa/test-cases/<rid>.md`**; `audit/perf-<rid>.md` only when the bug is performance-shaped (matches the L449-452 "When this applies" criteria) |
+> | config | (none) | `rd/security-review.md` only — this one is genuinely ridless (`SECURITY_REVIEW.relativePath` carries no `<rid>`) |
+>
+> The `<rid>`-bearing names are the ones `peaks request transition` resolves first; the pre-rid names (`rd/code-review.md`, `rd/security-review.md`, `audit/security.md`, `audit/perf.md`) stay **readable** as back-compat tiers for sessions written before slice `2026-09-14-audit-artifact-rid-scoping` — read, never written. Exception: `rd/perf-baseline.md` is the RD-side Gate B9 baseline (a different artifact from `audit/perf-<rid>.md`) and is still both written by `peaks perf baseline --apply` and read as `AUDIT_PERF`'s oldest legacy tier.
 > | docs / chore | (none) | (none) |
 >
 > The escape hatch `--allow-incomplete --reason "<text>"` still exists for one-off exceptions; the bypass is recorded in the artifact transition note.
@@ -68,16 +70,16 @@ npx vitest run --changed --reporter=verbose 2>&1 | tail -20
 
 **Peaks-Loop Gate B3 — Before QA handoff: code review evidence exists:**
 ```bash
-ls .peaks/_runtime/<sessionId>/rd/code-review.md 2>&1
-# Expected: .peaks/_runtime/<sessionId>/rd/code-review.md
+ls .peaks/_runtime/<sessionId>/rd/code-review-<rid>.md 2>&1
+# Expected: .peaks/_runtime/<sessionId>/rd/code-review-<rid>.md
 # "No such file" → BLOCKED. Run code review (use code-reviewer agent or equivalent),
 # record findings, fix CRITICAL/HIGH issues, then re-check.
 ```
 
 **Peaks-Loop Gate B4 — Before QA handoff: security review evidence exists:**
 ```bash
-ls .peaks/_runtime/<sessionId>/rd/security-review.md 2>&1
-# Expected: .peaks/_runtime/<sessionId>/rd/security-review.md
+ls .peaks/_runtime/<sessionId>/audit/security-<rid>.md 2>&1
+# Expected: .peaks/_runtime/<sessionId>/audit/security-<rid>.md
 # "No such file" → BLOCKED. Run security review (use security-reviewer agent or equivalent),
 # fix CRITICAL/HIGH issues, record findings, then re-check.
 ```

package/skills/bee/peaks-rd/references/writing-handoff-frontmatter.md

@@ -4,7 +4,7 @@ Every RD handoff artifact carries a **YAML frontmatter block** so peaks-qa (and
 
 ## Path
 
-`.peaks/_runtime/<sessionId>/prd/handoff.md` — the canonical immutable PRD handoff path (v2.11.0+). The handoff is written by peaks-prd, sha256-hashed in frontmatter, and verified by every downstream sub-agent against the dispatched hash before reading.
+`.peaks/_runtime/<sessionId>/prd/handoff-<rid>.md` — the canonical immutable PRD handoff path (v2.11.0+; one capsule per slice since `2026-09-14-prd-capsule-rid-scoping`). The pre-scoping `.peaks/_runtime/<sessionId>/prd/handoff.md` is still readable as the legacy tier. The handoff is written by peaks-prd, sha256-hashed in frontmatter, and verified by every downstream sub-agent against the dispatched hash before reading.
 
 > **v2.11.0 change (Group A):** the per-session `rd/tech-doc.md` is removed; the immutable peaks-prd handoff replaces it as the slice's source-of-truth architecture document.
 
@@ -19,7 +19,7 @@ scope:
 files:
   - src/services/slice/schema-router.ts
   - src/services/audit/audit-goal-service.ts
-handoffPath: .peaks/_runtime/<sessionId>/prd/handoff.md
+handoffPath: .peaks/_runtime/<sessionId>/prd/handoff-<rid>.md
 handoffHash: sha256:<64 hex chars>
 decisions:
   - id: D1
@@ -34,10 +34,10 @@ nextActions:
   - "If Gate C passes, transition to txt handoff"
 gateEvidence:
   projectScan: .peaks/project-scan/project-scan.md
-  prdHandoff: .peaks/_runtime/<sessionId>/prd/handoff.md
-  codeReview: .peaks/_runtime/<sessionId>/rd/code-review.md
-  securityReview: .peaks/_runtime/<sessionId>/rd/security-review.md
-  perfBaseline: .peaks/_runtime/<sessionId>/rd/perf-baseline.md
+  prdHandoff: .peaks/_runtime/<sessionId>/prd/handoff-<rid>.md
+  codeReview: .peaks/_runtime/<sessionId>/rd/code-review-<rid>.md
+  securityReview: .peaks/_runtime/<sessionId>/audit/security-<rid>.md
+  perfBaseline: .peaks/_runtime/<sessionId>/audit/perf-<rid>.md
 schemaVersion: '2.0'
 ---
 ```

package/skills/bee/peaks-reviewer/SKILL.md

@@ -4,7 +4,7 @@ visibility: internal
 description: Third-party independent code-review skill for Peaks. Use as a parallel reviewer to karpathy-reviewer when a workflow needs an out-of-band (non-RD-model) review pass to mitigate single-LLM self-dogfood blind spots. Emits a schema-validated ReviewerEnvelope (no free-form LLM JSON). v2.14.0 G4.
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/bee/peaks-sc/SKILL.md

@@ -8,7 +8,7 @@ description: |
 visibility: internal
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/bee/peaks-security-audit/SKILL.md

@@ -12,7 +12,7 @@ metadata:
     - schema: .peaks/project-scan/audit-output-schema.md
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/bee/peaks-txt/SKILL.md

@@ -8,7 +8,7 @@ description: |
 visibility: internal
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/bee/peaks-ui/SKILL.md

@@ -8,7 +8,7 @@ description: |
 visibility: internal
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/peaks-audit/SKILL.md

@@ -3,7 +3,7 @@ name: peaks-audit
 description: Gate autonomous LLM work on a 6-dimension audit + user-accepted goal. Use IMMEDIATELY after a need is expressed, BEFORE any PRD/RD/QA work starts.
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 

package/skills/peaks-code/SKILL.md

@@ -3,7 +3,7 @@ name: peaks-code
 description: Code-domain loop engineering orchestrator for the Peaks-Loop skill family. Use when the user asks Peaks-Loop to handle a code-repo workflow end-to-end (端到端/全流程/需求开发), especially from a product document (PRD/飞书文档/Feishu doc) through implementation and validation. Coordinates peaks-prd, peaks-rd, peaks-qa, peaks-ui, peaks-sc, and peaks-txt while preserving user confirmation gates. Triggers on `/peaks-code`, "peaks code", "全流程开发", "端到端迭代". General primitives (peaks-resume / peaks-status / peaks-test) are sibling skills, not children.
 ---
 
-<!-- peaks:loop-hygiene — generated block; keep byte-identical in every SKILL.md -->
+<!-- peaks:loop-hygiene — HAND-MAINTAINED block, NOT generated; byte-identical in every SKILL.md, enforced by tests/unit/skills/loop-hygiene-block.test.ts. Edit one, edit all 22. -->
 
 ## Loop hygiene (every turn — MANDATORY)
 
@@ -37,7 +37,7 @@ of ~30 minutes clears the gate's "already passed" list, so a file you cleared ea
 denied again after a long pause. That is the gate resetting, not you regressing.
 <!-- /peaks:loop-hygiene -->
 
-> **Detached sub-agent mode (Phase A, slice 2026-08-10).** When the orchestrator requires true parallelism with isolated context windows and survives orchestrator session exit, dispatch sub-agents with `--mode detached --vendor <claude|codex|copilot>`. The CLI spawns a real OS process via `ProcessSupervisor` (Windows `DETACHED_PROCESS` + `CREATE_NEW_PROCESS_GROUP`, POSIX `setsid` + `nohup`); the child vendor LLM receives a 5–8KB minimum prompt slice (no orchestrator session history) and self-compacts at 0.85 / 0.95 against the vendor window via the `<peaks-auto-compact>` marker (G8 — unlimited spend authorized). Orchestrator MUST emit one line of prose before every detached dispatch: `⏳ Spawning detached sub-agent via <vendor>: rid=<rid> (ETA ~60s)`. Status is read from `.peaks/_runtime/<sid>/detached/<rid>/status.json`; `LifecycleOwner` enforces 100% cleanup of `pid` / `log.txt` / `status.json` / `owner-session` on every exit path. `--no-throttle --max-concurrent <N>` bypasses `ResourceBudgetGuard` (user accepts risk; default max=8). See `references/sub-agent-dispatch.md` §"Detached Mode" for the full contract. Default mode remains `in-process` for backward compat (existing 106+ dispatch tests untouched).
+> **Detached sub-agent mode (Phase A, slice 2026-08-10).** When the orchestrator requires true parallelism with isolated context windows and survives orchestrator session exit, dispatch sub-agents with `--mode detached --vendor <claude|codex|copilot>`. The CLI spawns a real OS process via `ProcessSupervisor` (Windows `DETACHED_PROCESS` + `CREATE_NEW_PROCESS_GROUP`, POSIX `setsid` + `nohup`); the child vendor LLM receives a 5–8KB minimum prompt slice (no orchestrator session history) and, via the `<peaks-auto-compact>` marker, is told to persist durable state to `.peaks/_runtime/<sid>/detached/<rid>/compact/<n>.json` before its own harness compacts the session (G8, corrected 2026-09-13 rid `2026-09-13-defects-e`: the child cannot compact itself and nothing is spliced back into its prompt; peaks runtime simply has no token-cost throttle, which is not the same as a granted budget). Orchestrator MUST emit one line of prose before every detached dispatch: `⏳ Spawning detached sub-agent via <vendor>: rid=<rid> (ETA ~60s)`. Status is read from `.peaks/_runtime/<sid>/detached/<rid>/status.json`; `LifecycleOwner` enforces 100% cleanup of `pid` / `log.txt` / `status.json` / `owner-session` on every exit path. `--no-throttle --max-concurrent <N>` bypasses `ResourceBudgetGuard` (user accepts risk; default max=8). See `references/sub-agent-dispatch.md` §"Detached Mode" for the full contract. Default mode remains `in-process` for backward compat (existing 106+ dispatch tests untouched).
 ---
 
 ## Scope (RL-8 — red line, locked 2026-07-08)
@@ -203,7 +203,7 @@ Before the first planning action, run `peaks fresh-context preflight --prompt "<
 |---|---|---|
 | `< 0.85` | normal | skip — LLM keeps working |
 | `0.85 ≤ ratio < 0.95` | **pre-compact zone** | `peaks code auto-compact` fires **automatically** (deferred only when in-flight sub-agent batch is running; fires the moment the batch lands). The LLM does not prompt the user. |
-| `ratio ≥ 0.95` | **red-line (Karpathy §4)** | synchronous gate — `peaks code auto-compact` invoked immediately; `peaks code context-now` returns `action: 'red-line'` and refuses to advance until ratio drops below 0.85. **Karpathy §4 automatic exception** — LLM cannot opt out. |
+| `ratio ≥ 0.95` | **red-line (Karpathy §4)** | `peaks code auto-compact` invoked immediately; `peaks code context-now` returns `action: 'red-line'`. **Since 4.0.47 the red line REQUESTS the compaction and says it is waiting — it does NOT block sub-agent dispatch, and it does not refuse to advance.** Keep working and re-probe with `peaks code context-now`; the harness performs the compaction, and nothing peaks-loop can do lowers the ratio on its own, so blocking here was a deadlock rather than a gate. If the ratio keeps climbing and no compaction lands, report that and hand control back — do not stall. `--bypass-red-line` is a no-op. |
 
 **Probe primitive (single source of truth):** `peaks code context-now --json`. Do NOT use `peaks context check --prompt-size` (deprecated, will silently under-report ratio). Returns `{ ratio, action: 'ok' | 'soft-warn' | 'auto-compact-now' | 'red-line' }` — Code reads `action` and dispatches `peaks code auto-compact` on `auto-compact-now` or `red-line` without user confirmation.
 

package/skills/peaks-code/references/a2a-artifact-mapping.md

@@ -20,7 +20,7 @@ The mapping below uses peaks's own paths verbatim. Each row also notes where pea
 |---|---|---|---|
 | **AgentCard** (capability advertisement) | `peaks-skill-output-style` + `.peaks/.active-skill.json` | `.peaks/.active-skill.json`, `.peaks/.session.json` | peaks is a *local* tool, not a service. The "card" is the active-skill file plus a peek at `.peaks/PROJECT.md` for human-readable history. There is no `/.well-known/agent-card.json` endpoint. |
 | **Task** (stateful unit of work) | `peaks request` state machine for a single `<rid>` | `.peaks/_runtime/<sid>/{prd,rd,qa,ui,sc}/requests/<rid>.md` (the request artefact); `.peaks/_runtime/<sid>/<role>/session.json` (per-session metadata) | peaks's task lifecycle is `prd:confirmed-by-user → handed-off`, then per role `draft → spec-locked → implemented → qa-handoff`, then `qa:running → verdict-issued`. The full state graph is enforced by `peaks request transition`. A2A's Task object is JSON; peaks's task is **a set of files with a `state` field per role**. |
-| **Artifact** (immutable output) | `rd/tech-doc.md`, `rd/code-review.md`, `rd/security-review.md`, `qa/test-cases/<rid>.md`, `qa/test-reports/<rid>.md`, `qa/security-findings.md`, `qa/performance-findings.md`, `sc/handoff.md` | as listed | peaks's artefacts are *append-once*, not strictly immutable: a `qa/test-reports/<rid>.md` may be re-emitted on repair cycles. The convention is "newest write wins; the file at the end of the workflow is the truth", which is close enough to A2A's immutable-Artifact semantics for translation purposes. |
+| **Artifact** (immutable output) | `rd/tech-doc.md`, `rd/code-review-<rid>.md`, `audit/security-<rid>.md`, `qa/test-cases/<rid>.md`, `qa/test-reports/<rid>.md`, `qa/security-findings.md`, `qa/performance-findings.md`, `sc/handoff.md` | as listed | peaks's artefacts are *append-once*, not strictly immutable: a `qa/test-reports/<rid>.md` may be re-emitted on repair cycles. The convention is "newest write wins; the file at the end of the workflow is the truth", which is close enough to A2A's immutable-Artifact semantics for translation purposes. |
 | **Message** (non-artifact communication) | `peaks skill presence` heartbeat + transition `--reason` notes | `.peaks/.active-skill.json` (`lastHeartbeat`), transition notes in `.peaks/_runtime/<sid>/<role>/requests/<rid>.md` | peaks does **not** separate Messages from Artifacts at the storage layer; a "message" is anything that is not the artefact body (the `<!-- peaks-memory:start -->` markers, the `state` field, the `--reason` text on a transition). Treat these as inline metadata of the artefact, not as separate objects. |
 | **Part** (atomic content unit) | Markdown sections within an artefact, frontmatter fields | inline within the artefact | peaks's Artifacts are single Markdown files, so the "Part" concept maps to a heading or a frontmatter field. A `Part`'s `kind` in A2A terms is `text` (the prose), `file` (a `<!-- peaks-memory:start -->` block as a structured chunk), or `data` (the frontmatter). A2A's `form` / `iframe` / video `Part` kinds are not produced by peaks. |
 
@@ -78,8 +78,8 @@ A user runs `peaks-code` for a "add user authentication" feature. Mapping the re
 .peaks/_runtime/<sessionId>/ui/design-draft.md       → A2A Artifact (kind=visual-spec)
 .peaks/_runtime/<sessionId>/rd/tech-doc.md           → A2A Artifact (kind=implementation-plan)
 .peaks/_runtime/<sessionId>/qa/test-cases/001.md     → A2A Artifact (kind=test-cases)
-.peaks/_runtime/<sessionId>/rd/code-review.md        → A2A Artifact (kind=review, status=fixed)
-.peaks/_runtime/<sessionId>/rd/security-review.md    → A2A Artifact (kind=security-review)
+.peaks/_runtime/<sessionId>/rd/code-review-<rid>.md  → A2A Artifact (kind=review, status=fixed)
+.peaks/_runtime/<sessionId>/audit/security-<rid>.md  → A2A Artifact (kind=security-review)
 .peaks/_runtime/<sessionId>/qa/test-reports/001.md    → A2A Artifact (kind=test-report, verdict=pass)
 .peaks/_runtime/<sessionId>/qa/security-findings.md  → A2A Artifact (kind=security-findings)
 .peaks/_runtime/<sessionId>/qa/performance-findings.md → A2A Artifact (kind=performance-findings)

package/skills/peaks-code/references/local-artifact-workspace.md

@@ -53,7 +53,7 @@ Files written into these directories during the workflow (not pre-created — th
 - `rd/project-scan.md` (Code step 0.6)
 - `rd/tech-doc.md` (feature/refactor planning; required by `rd → implemented` gate)
 - `rd/bug-analysis.md` (bugfix planning; required by `rd → implemented` gate for `--type bugfix`)
-- `rd/code-review.md`, `rd/security-review.md` (required by `rd → qa-handoff` gate for feature/bugfix/refactor; security-review only for config)
+- `rd/code-review-<rid>.md`, `audit/security-<rid>.md` (required by `rd → qa-handoff` gate for feature/bugfix/refactor; the `config` type instead requires the ridless `rd/security-review.md`)
 - `rd/mock-plan.md` (frontend-only mode)
 - `ui/design-draft.md` (UI step)
 - `system/existing-system.md` (Code step 0.7; legacy projects only)

package/skills/peaks-code/references/resume-detection.md

@@ -12,12 +12,18 @@
 sid=$(peaks session info --active --project "$(git rev-parse --show-toplevel)" --json | python3 -c "import sys,json; print(json.load(sys.stdin)['data']['sessionId'])")
 
 # 2. Enumerate the session's artifact tree (one `find` call, no new CLI)
-find ".peaks/$sid/" -type f 2>/dev/null | sort
+find ".peaks/_runtime/$sid/" -type f 2>/dev/null | sort
 
 # 3. For each role request artifact present, read its `state:` field
-#    (one-pass grep; only files that exist)
-for f in .peaks/$sid/prd/requests/*.md .peaks/$sid/rd/requests/*.md .peaks/$sid/qa/requests/*.md; do
-  [ -f "$f" ] && echo "$f: $(grep -m1 '^state:' "$f" | awk '{print $2}')"
+#    (only files that exist). Artifacts are append-only: every round appends its
+#    own `## Status` block and `request transition` rewrites the NEWEST
+#    `- state:` line in place, so the LAST match is the current state and the
+#    earlier ones are history. Reading the FIRST match returns a superseded
+#    round — that was the 2026-09-14 defect. The leading `-` is part of the
+#    field: an unanchored `^state:` matches nothing (artifacts write
+#    `- state:`), so it reported an empty state, silently.
+for f in .peaks/_runtime/$sid/prd/requests/*.md .peaks/_runtime/$sid/rd/requests/*.md .peaks/_runtime/$sid/qa/requests/*.md; do
+  [ -f "$f" ] && echo "$f: $(grep '^-[[:space:]]*state:' "$f" | tail -n1 | sed 's/^-[[:space:]]*state:[[:space:]]*//; s/[[:space:]]*$//')"
 done
 
 # 4. Compute "deepest completed gate" by file-presence + state mapping
@@ -28,7 +34,7 @@ done
 
 | Files present | State | Deepest completed gate | Resume point (if any) |
 |---|---|---|---|
-| only `.peaks/$sid/.session.json` | (no slice) | (none) | fresh — skip to Step 1 |
+| only `.peaks/_runtime/$sid/session.json` | (no slice) | (none) | fresh — skip to Step 1 |
 | `prd/requests/<rid>.md` | `state: handed-off` | Gate B (swarm converged) | resume at Step 3 (swarm) — but if swarm already ran and produced `rd/tech-doc.md` / `qa/test-cases/<rid>.md`, drop to deepest |
 | `rd/requests/<rid>.md` | `state: qa-handoff` | Gate C (RD done) | resume at Step 6 (QA validation) |
 | `qa/requests/<rid>.md` | `state: verdict-issued` | Gate D (QA done) | resume at Step 10 (TXT handoff) |
@@ -39,8 +45,8 @@ done
 | Missing file | Resume at |
 |---|---|
 | `rd/tech-doc.md` (for `feature`/`refactor`) or `rd/bug-analysis.md` (for `bugfix`) | Step 3b (RD planning) |
-| `rd/code-review.md` or `rd/security-review.md` | Step 5 (RD review fan-out) |
-| `rd/perf-baseline.md` (for `feature`/`refactor`) | Step 5 (perf baseline) |
+| `rd/code-review-<rid>.md` or `audit/security-<rid>.md` | Step 5 (RD review fan-out) |
+| `audit/perf-<rid>.md` (for `feature`/`refactor`) | Step 5 (perf baseline) |
 | `qa/test-cases/<rid>.md` | Step 6 (QA test-case generation) |
 | `qa/test-reports/<rid>.md` or `qa/security-findings.md` or `qa/performance-findings.md` | Step 6 (QA execution) |
 | `txt/handoff.md` | Step 10 (TXT handoff) |

package/skills/peaks-code/references/runbook.md

@@ -153,8 +153,9 @@ peaks request transition <rid> --role rd --state implemented --project <repo> --
 
 # 5. Peaks-Loop Code review + security review BEFORE qa-handoff transition.
 #    Produce the evidence files the CLI gate enforces:
-#      - .peaks/_runtime/<id>/rd/code-review.md     (CRITICAL/HIGH findings + fixes; required for feature/bugfix/refactor)
-#      - .peaks/_runtime/<id>/rd/security-review.md (required for feature/bugfix/refactor/config)
+#      - .peaks/_runtime/<id>/rd/code-review-<rid>.md  (CRITICAL/HIGH findings + fixes; required for feature/bugfix/refactor)
+#      - .peaks/_runtime/<id>/audit/security-<rid>.md  (required for feature/bugfix/refactor; the ridless rd/security-review.md is the config-type requirement)
+#      The pre-rid names remain accepted back-compat tiers, but writing one re-opens the two-slices-one-session collision (slice 2026-09-14-audit-artifact-rid-scoping).
 #    Then transition. If --type is docs/chore the gate is empty and the transition is unguarded.
 peaks request transition <rid> --role rd --state qa-handoff --project <repo> --json
 

package/skills/peaks-code/references/session-overload-signal-index.md

@@ -45,7 +45,8 @@ LLM senses context pressure
   │   │     ├─ Job mode → MUST auto-compact now
   │   │     └─ non-Job  → wait for next idle turn, then compact
   │   ├─ 0.85 ≤ ratio < 0.95 → pre-compact zone; auto-compact MUST fire
-  │   └─ ratio ≥ 0.95       → red-line; synchronous gate (Karpathy §4)
+  │   └─ ratio ≥ 0.95       → red-line; compact REQUESTED, dispatch NOT blocked (Karpathy §4)
+  │                                 keep working, re-probe, do not stall
   ├─► Check statusline compact bar (signal #4)
   │   ├─ queued / preparing / compacting / verifying → wait
   │   ├─ failed → read failure reason, do NOT spawn new session

package/skills/peaks-code/references/sub-agent-dispatch.md

@@ -449,7 +449,7 @@ peaks skill presence:set peaks-code --project <repo> --mode <mode> --gate swarm-
 
 - **Cross-platform spawn**: Windows uses `DETACHED_PROCESS` + `CREATE_NEW_PROCESS_GROUP`; POSIX uses `setsid` + `nohup`. Implementation: `packages/peaks-loop-internal-runtime/src/process-supervisor.ts`.
 - **Minimum-context prompt**: `PromptBuilder` emits a 5–8KB slice `{rid, role, vendor, files, refs}` plus the verbatim `<peaks-auto-compact>` marker. The forbidden marker `@@@ORCHESTRATOR_SESSION_HISTORY_BOUNDARY@@@` MUST NOT appear in any prompt (unit-tested; regression fails vitest).
-- **G8 infinite context**: the child LLM self-monitors context; on ≥0.85 it compacts and writes `.peaks/_runtime/<sid>/detached/<rid>/compact/<n>.json`; on ≥0.95 it writes `status.json` note `'compact-emergency'`. peaks runtime does NOT throttle on token cost — user authorized unlimited spend for G8 (recorded as `tokenUsage` on the dispatch record for audit visibility only).
+- **G8 infinite context (corrected 2026-09-13, rid `2026-09-13-defects-e`)**: the child does **not** and **cannot** compact its own session. `/compact` is not model-invokable, hooks may only observe or veto, there is no `--compact` flag, and no external process can reach a running session's memory — compaction is the child's own harness acting on its own window. What the `<peaks-auto-compact>` marker actually asks for is the reachable half: write the durable state (summary / task state / decisions kept / options discarded) to `.peaks/_runtime/<sid>/detached/<rid>/compact/<n>.json` and heartbeat to `status.json`, so nothing load-bearing lives only in a conversation that will be summarised. Nothing flows back into the child's context — its prompt is the `-p` argv string fixed at spawn. Token cost: peaks runtime has no token-cost throttle to relax (the only guard, `ResourceBudgetGuard`, throttles RSS / CPU / fan-out, never spend), so there is no cost cap — but that is an absence of throttling, not a granted permission with a flag behind it.
 - **Lifecycle closure invariant**: `pid` / `log.txt` / `status.json` / `owner-session` 100% cleaned on every exit path (success / crash / OOM / SIGTERM). `peaks sub-agent cleanup --orphan` is the only orphan-killing path (RL-15: user-only decision; default is "do nothing").
 - **Resource budget guard**: `ResourceBudgetGuard` enforces `runtime RSS ≤ 200MB / idle CPU ≤ 5% / fan-out ≤ 8 (default)`. `--no-throttle` bypasses (user accepts risk; adds warning to envelope). `--max-concurrent <N>` overrides the 8 default.
 - **Vendor-neutral**: `VendorAdapterRegistry` registers `ClaudeAdapter` (Phase A) / `CodexAdapter` + `CopilotAdapter` (Phase B). Missing vendor CLI → fallback to default with `dispatchRecord.warning = 'vendor fallback: <id>→claude'`.

package/skills/peaks-code/references/workflow-gates-and-types.md

@@ -104,11 +104,13 @@ The CLI gate (`peaks request transition --state qa-handoff`) is the authoritativ
 
 | Request type | Required RD evidence (under `.peaks/_runtime/<sessionId>/`) |
 |---|---|
-| `feature` / `refactor` | `rd/tech-doc.md` + `rd/code-review.md` + `rd/security-review.md` + `rd/perf-baseline.md` + `qa/test-cases/<rid>.md` (qa/test-cases pre-drafted by the 4th sub-agent in peaks-rd's parallel fan-out — slice 004) |
-| `bugfix` | `rd/bug-analysis.md` + `rd/code-review.md` + `rd/security-review.md` + `qa/test-cases/<rid>.md` (rd/perf-baseline.md only when the bug is performance-shaped) |
-| `config` | `rd/security-review.md` |
+| `feature` / `refactor` | `rd/tech-doc.md` + `rd/code-review-<rid>.md` + `audit/security-<rid>.md` + `audit/perf-<rid>.md` + `qa/test-cases/<rid>.md` (qa/test-cases pre-drafted by the 4th sub-agent in peaks-rd's parallel fan-out — slice 004) |
+| `bugfix` | `rd/bug-analysis.md` + `rd/code-review-<rid>.md` + `audit/security-<rid>.md` + `audit/perf-<rid>.md` + `qa/test-cases/<rid>.md` (the perf artifact is always required; a bug with no perf surface satisfies it with the `N/A — no perf surface` stub) |
+| `config` | `rd/security-review.md` (this one is genuinely ridless — `SECURITY_REVIEW.relativePath` carries no `<rid>`) |
 | `docs` / `chore` | (no extra evidence required) |
 
+> The four rid-scoped names above are the gate's **primary** tier (slice `2026-09-14-audit-artifact-rid-scoping`): two slices in one session share `rd/` and `audit/`, so a ridless write silently destroyed the earlier slice's evidence. The pre-rid names (`rd/code-review.md`, `audit/security.md`, `audit/perf.md`, `rd/security-review.md`, `rd/perf-baseline.md`) remain accepted as **back-compat tiers** — they keep old sessions passing, but a new slice must write the rid-scoped name or it reopens the collision.
+
 Always required (in addition to the type-specific row): `ls .peaks/_runtime/<sessionId>/rd/requests/<rid>.md`. Missing any required file → DO NOT attempt the qa-handoff transition; CLI will reject with PREREQUISITES_MISSING.
 
 ```bash
@@ -116,10 +118,10 @@ Always required (in addition to the type-specific row): `ls .peaks/_runtime/<ses
 ls .peaks/_runtime/<sessionId>/rd/requests/<rid>.md
 
 # Type-specific RD evidence (must match the type recorded in the artifact body)
-#   feature / refactor → ls rd/tech-doc.md rd/code-review.md rd/security-review.md rd/perf-baseline.md qa/test-cases/<rid>.md
+#   feature / refactor → ls rd/tech-doc.md rd/code-review-<rid>.md audit/security-<rid>.md audit/perf-<rid>.md qa/test-cases/<rid>.md
 #                         (qa/test-cases/<rid>.md pre-drafted by the 4th sub-agent in peaks-rd's parallel fan-out — slice 004)
-#   bugfix             → ls rd/bug-analysis.md rd/code-review.md rd/security-review.md qa/test-cases/<rid>.md
-#                         (rd/perf-baseline.md only when the bug is performance-shaped)
+#   bugfix             → ls rd/bug-analysis.md rd/code-review-<rid>.md audit/security-<rid>.md audit/perf-<rid>.md qa/test-cases/<rid>.md
+#                         (the perf artifact is always required; use the "N/A — no perf surface" stub when there is no perf surface)
 #   config             → ls rd/security-review.md
 #   docs / chore       → (no extra evidence required)
 # Missing any required file → DO NOT attempt the qa-handoff transition; CLI will reject with PREREQUISITES_MISSING.