@mmerterden/multi-agent-pipeline 13.5.0 → 14.0.0

package/pipeline/commands/multi-agent/SKILL.md

@@ -6,7 +6,7 @@ allowed-tools: Agent, Bash, Read, Write, Edit, Glob, Grep, TaskCreate, TaskUpdat
 
 # Multi-Agent Task Orchestrator
 
-> **MUST: Figma MCP is BLOCKING for any task with a Figma reference.** Before any UI synthesis runs in this pipeline (analysis, planning, dev, or rework), call `mcp__claude_ai_Figma__get_design_context` for every referenced frame and use the `CodeConnectSnippet` component name verbatim. Backend-only tasks are exempt. See `$HOME/.claude/multi-agent-refs/rules.md` "User Interaction Discipline" and `pipeline/rules/figma-pipeline.md` "MUST: Figma MCP-first (BLOCKING)".
+> **MUST: Figma MCP is BLOCKING for any task with a Figma reference.** Before any UI synthesis runs in this pipeline (analysis, planning, dev, or rework), call `mcp__claude_ai_Figma__get_design_context` for every referenced frame and use the `CodeConnectSnippet` component name verbatim. Backend-only tasks are exempt. See `$HOME/.claude/multi-agent-refs/rules.md` "User Interaction Discipline" and `$HOME/.claude/rules/figma-pipeline.md` "MUST: Figma MCP-first (BLOCKING)".
 
 Parse the user input and route to the correct sub-command.
 
@@ -87,8 +87,8 @@ Lib scripts (`~/.claude/lib/`):
 | `stack [ios\|android\|backend\|mobile\|all]` | Swap skills for next conversation. No arg = show current stack. |
 | `language [en\|tr]` | Show or set the assistant `outputLanguage` (explanations and chat replies). `promptLanguage` is locked to `en` and is not toggleable. No arg = show current `outputLanguage`. With `en` or `tr` = set and persist `outputLanguage`. External payloads (commits, PR bodies, Jira) stay English. |
 | `setup` | Keychain token + Git Identity onboarding |
-| `--dev` | Dev-only: Init -> Dev(Opus) -> Commit -> Report |
-| `--dev autopilot` or `dev-autopilot` | Fastest path  -  Dev(Opus) + zero interaction |
+| `--dev` | Dev-only: Init -> Dev(Opus) -> Review -> Test -> Commit -> Report |
+| `--dev autopilot` or `dev-autopilot` | Fastest path  -  Dev(Opus) + Review (auto-fix) + zero interaction |
 | `--local` | No worktree  -  works directly on local branch |
 | `autopilot` | Skip user confirmations, auto commit/PR |
 | No args / `help` | Show usage guide |
@@ -121,6 +121,8 @@ This command uses lazy loading for token efficiency. Read the relevant sub-file
 | `build-optimize` | `$HOME/.claude/commands/multi-agent/build-optimize/SKILL.md` |
 | `local` | `$HOME/.claude/commands/multi-agent/local/SKILL.md` |
 | `local-autopilot` | `$HOME/.claude/commands/multi-agent/local-autopilot/SKILL.md` |
+| `dev` | `$HOME/.claude/commands/multi-agent/dev/SKILL.md` |
+| `dev-autopilot` | `$HOME/.claude/commands/multi-agent/dev-autopilot/SKILL.md` |
 | `dev-local` | `$HOME/.claude/commands/multi-agent/dev-local/SKILL.md` |
 | `dev-local-autopilot` | `$HOME/.claude/commands/multi-agent/dev-local-autopilot/SKILL.md` |
 | `create-jira` | `$HOME/.claude/commands/multi-agent/create-jira/SKILL.md` (loads `$HOME/.claude/multi-agent-refs/generate-issue.md`) |
@@ -162,7 +164,7 @@ This command uses lazy loading for token efficiency. Read the relevant sub-file
 
 ## language Command
 
-Toggles only `outputLanguage` (`promptLanguage` is locked to `"en"`  -  see `$HOME/.claude/multi-agent-refs/rules.md` section "Language Application"). Full spec: `pipeline/commands/multi-agent/language/SKILL.md`.
+Toggles only `outputLanguage` (`promptLanguage` is locked to `"en"`  -  see `$HOME/.claude/multi-agent-refs/rules.md` section "Language Application"). Full spec: `$HOME/.claude/commands/multi-agent/language/SKILL.md`.
 
 ---
 
@@ -229,7 +231,7 @@ Save to `prefs.projects[{project}].branches`.
 ```
 Pipeline mode:
   1. Full pipeline (8 phases, Sonnet dev, parallel review + triage  -  CLI-aware reviewer set)
-  2. --dev (fast: Init → Dev(Opus) → Commit → Report)
+  2. --dev (fast: Init → Dev(Opus) → Review → Test → Commit → Report  -  no analysis/planning)
 Select [1/2]:
 ```
 

package/pipeline/commands/multi-agent/analysis/SKILL.md

@@ -45,7 +45,7 @@ When citing a Locked decision in code or docs, prefer `Locked <n> (<short label>
 9. **Per-platform output split.** One markdown file is produced per selected platform under `analysis/<feature>-<platform>.md`. There is no merged single file. Section duplication follows the A3 hybrid table in `$HOME/.claude/multi-agent-refs/analysis-template.md` (Sections 1 + 4 duplicated verbatim; Sections 2, 3, 5, 6, 7 projected per platform). Each file carries a YAML front-matter header naming the feature, platform, generated-at timestamp, sibling file list, and binding standards source.
 10. **Output destination is asked after drafts exist.** Phase 3 renders each per-platform markdown to `/tmp/analysis-<feature-slug>-<timestamp>/` first. Only then does Phase 3.5 surface the output-destination picker (Local / Confluence / Jira). Drafts in `/tmp/` are the source of truth on resume; `/multi-agent:resume` re-uses them when `phase == awaiting_output_decision`.
 11. **Repo-evidence reuse-first.** Phase 1b runs a repo-evidence collector against each selected repo, producing 13 buckets (services, dtos, useCases, validationRules, domainEntities, routes, coordinators, diConfigurators, uiComponents, tokens, localizationKeys, testingIdentifiers, analyticsEvents) with each row tagged `direct-match | same-domain | cross-cutting`. Section 7 emits `Reuse existing X (file:line)` rows for `direct-match` items and advisory rows for `cross-cutting` items. New-write tasks for items that have a `direct-match` are downgraded to Risks ("existing X candidate found; reuse or document why a new one is needed").
-12. **MUST: Figma access - 3-tier fallback chain, pipeline-wide.** When any Figma URL or node ID is supplied (Phase 0 Step 5), Phase 1 establishes a Figma ground-truth artefact via the 3-tier chain before any UI synthesis runs: (1) Figma MCP `get_design_context` / `get_screenshot` / `get_metadata` for every referenced frame, with one re-auth retry on auth failure; (2) Figma REST API (`GET /v1/files/{fileKey}/nodes`, `GET /v1/images/{fileKey}`) using the PAT resolved through `~/.claude/lib/credential-store.sh get <logical-key>` where `<logical-key>` = `prefs.global.keychainMapping.figma_pat`; (3) a user-attached screenshot as last resort. The tier in use is persisted as `state.figmaAccess.tier`. On Tier 1 the `CodeConnectSnippet` blocks name the exact target component consumed verbatim in Section 2 and Section 7. On Tier 2 the canonical-component decision falls back to the repo's `*.figma.swift` / `*.figma.kt` mappings keyed by `fileKey` + `nodeId`. On Tier 3 the canonical-component decision becomes "tier-3 best-fit pending design review" and produces a forced Open Question in Section 7 plus a Phase 4 `review_blocking` flag. Sound-alike alternatives, "more flexible" wrappers, and extrapolating from Confluence text-only "Component Kompozisyonu" / "Component Inventory" tables are forbidden in every tier. If all three tiers fail, halt the run and ask the user for access; never proceed with text-derived guesses. Section 7 architecture decisions must cite the node ID (Tier 1 or Tier 2) or the user-screenshot reference (Tier 3) for every UI atom row. Violations cost rebuild rounds (raw-primitive substitution; sound-alike-component swap). Generic rule rationale and checklist: see `pipeline/rules/figma-pipeline.md` "MUST: Figma access - 3-tier fallback chain (BLOCKING, pipeline-wide)". Memory: `[[figma-no-guesswork]]`.
+12. **MUST: Figma access - 3-tier fallback chain, pipeline-wide.** When any Figma URL or node ID is supplied (Phase 0 Step 5), Phase 1 establishes a Figma ground-truth artefact via the 3-tier chain before any UI synthesis runs: (1) Figma MCP `get_design_context` / `get_screenshot` / `get_metadata` for every referenced frame, with one re-auth retry on auth failure; (2) Figma REST API (`GET /v1/files/{fileKey}/nodes`, `GET /v1/images/{fileKey}`) using the PAT resolved through `~/.claude/lib/credential-store.sh get <logical-key>` where `<logical-key>` = `prefs.global.keychainMapping.figma`; (3) a user-attached screenshot as last resort. The tier in use is persisted as `state.figmaAccess.tier`. On Tier 1 the `CodeConnectSnippet` blocks name the exact target component consumed verbatim in Section 2 and Section 7. On Tier 2 the canonical-component decision falls back to the repo's `*.figma.swift` / `*.figma.kt` mappings keyed by `fileKey` + `nodeId`. On Tier 3 the canonical-component decision becomes "tier-3 best-fit pending design review" and produces a forced Open Question in Section 7 plus a Phase 4 `review_blocking` flag. Sound-alike alternatives, "more flexible" wrappers, and extrapolating from Confluence text-only "Component Kompozisyonu" / "Component Inventory" tables are forbidden in every tier. If all three tiers fail, halt the run and ask the user for access; never proceed with text-derived guesses. Section 7 architecture decisions must cite the node ID (Tier 1 or Tier 2) or the user-screenshot reference (Tier 3) for every UI atom row. Violations cost rebuild rounds (raw-primitive substitution; sound-alike-component swap). Generic rule rationale and checklist: see `$HOME/.claude/rules/figma-pipeline.md` "MUST: Figma access - 3-tier fallback chain (BLOCKING, pipeline-wide)". Memory: `[[figma-no-guesswork]]`.
 13. **Gherkin user stories.** Section 4 scenarios use Given / When / Then. Plain-prose user stories are rejected at render time.
 14. **Goals and Non-Goals are paired.** Section 2 always carries both columns. A row in only the goal column without an explicit non-goal counterpart is rejected. Vague "out of scope" phrasing does not satisfy the non-goal column; each non-goal names what is excluded.
 15. **New assets default to SVG.** Section 8 entries marked `new` are SVG unless a documented exception is captured in the rationale column (Lottie for motion design, optimized PNG for raster-only icons). PDF, JPG, and unoptimized PNG are rejected.
@@ -65,7 +65,7 @@ When citing a Locked decision in code or docs, prefer `Locked <n> (<short label>
 27. **Evidence digest caches Phase 1b and 1c.** `evidence_digest = sha256(featureName || sorted(platforms) || repoEvidence.summary || conventions.summary)`. When the same feature name is invoked again against the same set of repos and the digest matches, Phase 1b and 1c are skipped and the cached `evidence.repoEvidence` / `evidence.conventions` is reused. Cache TTL is 24 hours; manual invalidation via `--no-cache` flag.
 28. **SwiftUI Preview block mandatory (iOS projection, SwiftUI only).** When the iOS file is produced AND the affected view is a SwiftUI view (detected via `import SwiftUI` + `: View` protocol conformance in `evidence.repoEvidence[<repo>].buckets.uiComponents`), Section 13.6 renders a Preview block table covering at minimum: canonical default (LTR Light), Dark, RTL, Dynamic Type accessibilityLarge, and one error variant. Loading state and edge-case variants are added when distinct from canonical. UIKit-only features (no SwiftUI view artefact) drop Section 13.6 with note `(N/A: UIKit-only feature)`. Preview macro convention (`#Preview` for Swift 5.9+ vs legacy `PreviewProvider`) is read from `evidence.conventions[<repo>].previewMacro`. Each Preview variant listed in Section 13.6 must have a matching row in Section 15.2 Snapshot Tests; a Preview without a snapshot row triggers a Section 20 Risk.
 29. **Variant usage explicit and bounded.** Section 6 inventory rows list which variants this feature consumes per component (concrete enum case + bool value). New Section 6.X (Variant Usage Matrix) catalogues the full variant axis vs. used subset with a rationale per excluded variant. Sections 13.6 (Preview) and 15.2 (Snapshot) cover only the used subset; expanding the variant set requires updating Section 6.X first.
-30. **Analysis as self-contained design bridge - no MCP outside analysis phase (BLOCKING, pipeline-wide).** The analysis document is the sole design source for every downstream phase. After Phase 1 of `/multi-agent:analysis` produces `analysis/<feature>-<platform>.md`, Phase 2 Planning, Phase 3 Dev, Phase 4 Review, Phase 5 Test, Phase 6 Commit, and Phase 7 Report consume only the analysis document plus repo Code Connect mappings (`*.figma.swift` / `*.figma.kt`). Calling `mcp__claude_ai_Figma__*`, hitting `api.figma.com`, or fetching a `figma.com/design/...` URL during Phase 2+ is a violation. Applies to every mode that runs Phase 2+: `/multi-agent`, `/multi-agent:autopilot`, `/multi-agent:local`, `/multi-agent:local-autopilot`, `/multi-agent:dev`, `/multi-agent:dev-autopilot`, `/multi-agent:dev-local`, `/multi-agent:dev-local-autopilot`. Hard requirement (v9.0.0): Phase 2 Pre-item and Phase 3 Pre-item (BLOCKING) abort the run when the analysis document is missing. Memory: `[[mcp-only-in-analysis]]`. Generic rule rationale and access matrix: see `pipeline/rules/figma-pipeline.md` "MUST: No MCP outside analysis phase".
+30. **Analysis as self-contained design bridge - no MCP outside analysis phase (BLOCKING, pipeline-wide).** The analysis document is the sole design source for every downstream phase. After Phase 1 of `/multi-agent:analysis` produces `analysis/<feature>-<platform>.md`, Phase 2 Planning, Phase 3 Dev, Phase 4 Review, Phase 5 Test, Phase 6 Commit, and Phase 7 Report consume only the analysis document plus repo Code Connect mappings (`*.figma.swift` / `*.figma.kt`). Calling `mcp__claude_ai_Figma__*`, hitting `api.figma.com`, or fetching a `figma.com/design/...` URL during Phase 2+ is a violation. Applies to every mode that runs Phase 2+: `/multi-agent`, `/multi-agent:autopilot`, `/multi-agent:local`, `/multi-agent:local-autopilot`, `/multi-agent:dev`, `/multi-agent:dev-autopilot`, `/multi-agent:dev-local`, `/multi-agent:dev-local-autopilot`. Hard requirement (v9.0.0): Phase 2 Pre-item and Phase 3 Pre-item (BLOCKING) abort the run when the analysis document is missing. Memory: `[[mcp-only-in-analysis]]`. Generic rule rationale and access matrix: see `$HOME/.claude/rules/figma-pipeline.md` "MUST: No MCP outside analysis phase".
 31. **Business-rule to acceptance-criterion to test traceability (AI + human spine).** The analysis is a development handoff that both an AI implementer and a human reviewer must act on, so it is bound by one shared-ID vocabulary. Every business rule carries a stable id `BR-<slug>-NN` (Section 4.4). Each rule maps to at least one acceptance criterion written Given / When / Then (binary - two readers must not be able to disagree on pass/fail). Each acceptance criterion maps to unit-test scenarios in Section 15.1, one row per case across happy / boundary / error / empty-nil (enumerate at least the failure modes; agents hallucinate error handling when it is omitted). The same ids thread onward: Section 15.6 UI-test flows reference the `BR-` ids and use stable selectors (accessibilityIdentifier / testTag), Section 16 accessibility items reuse those identifiers, Section 11 analytics events cite their triggering rule or story, and Section 5/7 layout cells carry token + Figma node refs. Never invent copy or values (blank beats a guess; a missing source becomes a Section 20 Open Question). **Mode-aware gate:** in Full mode a business rule with no acceptance criterion, or an acceptance criterion with no Section 15.1 scenario, fails the dispatch gate. In **Lite mode Section 15 is not rendered**, so the rule-to-test half does not apply - Section 4.4 still lists each rule with its Given/When/Then acceptance criterion (the acceptance criterion is itself the testable statement), and the 15.1 mapping is deferred to whenever the feature is later analyzed in Full or implemented via `/multi-agent:dev`. The rule-to-acceptance-criterion half always holds, in both modes.
 
 ## Input
@@ -236,7 +236,7 @@ For each entry in `state.analysisSpec.contextLinks[]`, fan out by `type`. Entrie
 |------|---------|---------------|
 | swagger | `~/.claude/lib/fetch-swagger.sh <url>` | `state.analysisSpec.evidence.swagger[]` |
 | confluence | `~/.claude/lib/fetch-confluence.sh <url>` | `state.analysisSpec.evidence.confluence[]` |
-| figma | **MUST** establish the Figma access tier per Locked decision #12. Tier 1 (preferred): call `mcp__claude_ai_Figma__get_design_context` (primary), `mcp__claude_ai_Figma__get_screenshot`, `mcp__claude_ai_Figma__get_metadata` for every node ID; auth failure runs `authenticate` + `complete_authentication` then retries. Tier 2 (fallback): `GET https://api.figma.com/v1/files/{fileKey}/nodes?ids={nodeId}` for design context and `GET /v1/images/{fileKey}?ids={nodeId}&format=png&scale=2` for screenshots, with PAT from `~/.claude/lib/credential-store.sh get <logical-key>` (logical key = `prefs.global.keychainMapping.figma_pat`); canonical component resolves via repo `*.figma.swift` / `*.figma.kt` mapping. Tier 3 (last resort): user-attached screenshot, `codeConnectSnippets: []`, forced Open Question. Capture each `CodeConnectSnippet` block verbatim when on Tier 1 (component name + modifier chain). **Annotation capture (when the project `figma-config` has `annotations.enabled`):** on Tier 1, walk the `annotations[]` the `get_design_context` payload returns for each node (read `label` / `labelMarkdown`); on Tier 2, run `~/.claude/lib/fetch-figma-annotations.sh --file-key <fileKey> --node-id <ids> --lang-prefixes <config.annotations.langPrefixes>`. Parse each annotation by the configured language prefixes (prefixed `TR:/EN:` wins, else line1/line2, else single) and persist to `annotations[]`. The annotation is the authoritative copy for its node (Locked 3); the visible text layer is a placeholder. Persist `state.figmaAccess.tier`. See Locked decision #12 and `pipeline/rules/figma-pipeline.md` "MUST: Figma access - 3-tier fallback chain". | `state.analysisSpec.evidence.figma[]` (records: `nodeId`, `screenshotUrl`, `codeConnectSnippets[]`, `tokens[]`, `textLayers[]`, `annotations[]`, `tier`; see `analysis-spec.schema.json`) |
+| figma | **MUST** establish the Figma access tier per Locked decision #12. Tier 1 (preferred): call `mcp__claude_ai_Figma__get_design_context` (primary), `mcp__claude_ai_Figma__get_screenshot`, `mcp__claude_ai_Figma__get_metadata` for every node ID; auth failure runs `authenticate` + `complete_authentication` then retries. Tier 2 (fallback): `GET https://api.figma.com/v1/files/{fileKey}/nodes?ids={nodeId}` for design context and `GET /v1/images/{fileKey}?ids={nodeId}&format=png&scale=2` for screenshots, with PAT from `~/.claude/lib/credential-store.sh get <logical-key>` (logical key = `prefs.global.keychainMapping.figma`); canonical component resolves via repo `*.figma.swift` / `*.figma.kt` mapping. Tier 3 (last resort): user-attached screenshot, `codeConnectSnippets: []`, forced Open Question. Capture each `CodeConnectSnippet` block verbatim when on Tier 1 (component name + modifier chain). **Annotation capture (when the project `figma-config` has `annotations.enabled`):** on Tier 1, walk the `annotations[]` the `get_design_context` payload returns for each node (read `label` / `labelMarkdown`); on Tier 2, run `~/.claude/lib/fetch-figma-annotations.sh --file-key <fileKey> --node-id <ids> --lang-prefixes <config.annotations.langPrefixes>`. Parse each annotation by the configured language prefixes (prefixed `TR:/EN:` wins, else line1/line2, else single) and persist to `annotations[]`. The annotation is the authoritative copy for its node (Locked 3); the visible text layer is a placeholder. Persist `state.figmaAccess.tier`. See Locked decision #12 and `$HOME/.claude/rules/figma-pipeline.md` "MUST: Figma access - 3-tier fallback chain". | `state.analysisSpec.evidence.figma[]` (records: `nodeId`, `screenshotUrl`, `codeConnectSnippets[]`, `tokens[]`, `textLayers[]`, `annotations[]`, `tier`; see `analysis-spec.schema.json`) |
 | jira | `gh api` or Jira REST API for issue summary | `state.analysisSpec.evidence.jira[]` |
 | generic-doc | WebFetch on demand | `state.analysisSpec.evidence.confluence[]` (generic bucket) unless `binding: true` -> `evidence.standards[]` |
 | `local-file` | `Read` (no fetch) on tilde-expanded absolute path | `state.analysisSpec.evidence.standards[]` |
@@ -431,7 +431,7 @@ synthesizedSections = {
 | 3. Localization | If both Figma annotations/text layers and `evidence.repoEvidence[*].buckets.localizationKeys` are empty → `byPlatform[*]` = null. Otherwise produce the key table per the project `localization.ownership` mode (Locked 20): `in-repo` fills the configured locale set (default `ar, de, en, es, fr, it, ru, tr`, RTL for ar); `externally-owned` lists key + status + copy source + base value and defers per-locale values. Base copy is sourced from the Figma annotation when present (Locked 3). |
 | 4. API Contracts | If `evidence.swagger[]` is empty AND no Confluence embedded API table AND no `evidence.repoEvidence[*].buckets.services` direct-match → null. Otherwise produce endpoint summary + per-endpoint tables (shared verbatim across files). |
 | 5. Deeplink / Push | Per-platform: iOS uses Universal Links / `UNUserNotificationCenter`; Android uses `intent-filter` / FCM; Frontend uses web URL routing; Backend file omits this section entirely. |
-| 6. Business + Tests | If sections 2 and 4 are both `null` AND `evidence.firebase[]` is empty → null. Otherwise produce use-case + mock stubs + per-platform test skeletons + shared Firebase events table. Reuse Red-Green-Refactor naming from `pipeline/rules/tdd.md`. |
+| 6. Business + Tests | If sections 2 and 4 are both `null` AND `evidence.firebase[]` is empty → null. Otherwise produce use-case + mock stubs + per-platform test skeletons + shared Firebase events table. Reuse Red-Green-Refactor naming from `$HOME/.claude/rules/tdd.md`. |
 | 7. Development Plan | Always present. Tasks for the current platform only. Architecture standards come from `evidence.standards[]` filtered by platform (see Pass B step 1). **Reuse-first rule (Locked 11)**: when an item has `direct-match` in `evidence.repoEvidence[<repo>].buckets.<X>`, emit `Reuse existing <FQN> (<file>:<line>)` instead of `Add new <FQN>`. New-write task with a `direct-match` competitor becomes a Risk row. |
 
 #### Phase 2a - Pass B preview (Locked 26)
@@ -495,7 +495,7 @@ For each `platform` in `state.analysisSpec.platforms[]`:
 6. **Concatenate non-null sections in canonical order.** Numbering stays sequential `1..N` over the rendered set (omitted sections do not create gaps).
 7. **Schema validation** on the per-platform spec object:
    ```bash
-   python3 -c "import json,jsonschema; jsonschema.validate(json.load(open('state/<feature>-<platform>.json')), json.load(open('pipeline/schemas/analysis-spec.schema.json')))"
+   python3 -c "import json,jsonschema; jsonschema.validate(json.load(open('state/<feature>-<platform>.json')), json.load(open('$HOME/.claude/schemas/analysis-spec.schema.json')))"
    ```
    On failure for any platform, surface the error and stop before Phase 3 (do not draft partial outputs).
 
@@ -662,11 +662,11 @@ When `phase == "cancelled_at_pass_b_preview"`:
 | `~/.claude/lib/extract-conventions.sh` | Phase 1c convention extractor (7 pattern groups, JSON output, confidence levels) |
 | `~/.claude/lib/figma-screenshot.sh` | Phase 2b Tier 2 Figma image downloader (REST API, section drill, 2x scale PNG, manifest.json) |
 | `~/.claude/lib/md2confluence-v3.py` | Phase 4 Confluence dispatch (multipart attachments, `<ac:image>` injection, mermaid macro + fallback, tooltip footnote macro, punctuation gate) |
-| `pipeline/skills/shared/external/humanizer/SKILL.md` | Phase 3 tone pass |
+| `$HOME/.claude/skills/humanizer/SKILL.md` | Phase 3 tone pass |
 | 8-locale set (ar, de, en, es, fr, it, ru, tr) + localization-key naming (inline) | Section 10 localization generation |
 | `$HOME/.claude/multi-agent-refs/channels/confluence.md` | Phase 4 Confluence dispatch |
 | `$HOME/.claude/multi-agent-refs/channels/jira.md` | Phase 4 Jira dispatch |
-| `pipeline/rules/tdd.md` | Section 15 test naming |
+| `$HOME/.claude/rules/tdd.md` | Section 15 test naming |
 | `$HOME/.claude/multi-agent-refs/analysis-template.md` | Template master copy + language matrix (v3 - 23 main sections + 3 footer) |
 | `$HOME/.claude/multi-agent-refs/conventions-defaults.md` | Pass B fallback defaults (4 platforms x 7 pattern groups) - applied when convention confidence is none AND standards binding is silent |
 | `$HOME/.claude/scripts/validate-analysis-doc.mjs` | Phase 4 pre-dispatch gate: deterministic check of the emitted per-platform doc (front-matter, never-omitted sections, humanizer punctuation, Full-mode BR traceability) |

package/pipeline/commands/multi-agent/analysis-resolve/SKILL.md

@@ -121,7 +121,7 @@ For each open row in source order:
 | `$HOME/.claude/multi-agent-refs/analysis-template.md` | Section 20 / Section 23 table contracts, front-matter shape, omission re-flow |
 | `$HOME/.claude/multi-agent-refs/conventions-defaults.md` | `convention-fallback` default values per platform |
 | `~/.claude/lib/extract-conventions.sh` | fresh single-field extraction for convention-fallback candidates |
-| `pipeline/skills/shared/external/humanizer/SKILL.md` | tone reference for longer merged fragments (short fragments only need the punctuation gate) |
+| `$HOME/.claude/skills/humanizer/SKILL.md` | tone reference for longer merged fragments (short fragments only need the punctuation gate) |
 
 ## Notes
 

package/pipeline/commands/multi-agent/build-optimize/SKILL.md

@@ -62,13 +62,13 @@ Before dispatch, fail fast on non-iOS repos. Detect iOS context via the standard
 
 | Path | Reason |
 |------|--------|
-| `pipeline/skills/shared/external/xcode-build-orchestrator/SKILL.md` | The orchestrator this wrapper dispatches to |
-| `pipeline/skills/shared/external/xcode-build-benchmark/SKILL.md` | Baseline timing |
-| `pipeline/skills/shared/external/xcode-compilation-analyzer/SKILL.md` | Swift compile hotspots |
-| `pipeline/skills/shared/external/xcode-project-analyzer/SKILL.md` | Build settings / script phases / parallelism |
-| `pipeline/skills/shared/external/spm-build-analysis/SKILL.md` | SPM graph + plugins |
-| `pipeline/skills/shared/external/xcode-build-fixer/SKILL.md` | Apply approved fixes + re-benchmark |
-| `pipeline/skills/shared/external/NOTICE-xcode-build-skills.md` | MIT attribution + upstream pin |
+| `$HOME/.claude/skills/xcode-build-orchestrator/SKILL.md` | The orchestrator this wrapper dispatches to |
+| `$HOME/.claude/skills/xcode-build-benchmark/SKILL.md` | Baseline timing |
+| `$HOME/.claude/skills/xcode-compilation-analyzer/SKILL.md` | Swift compile hotspots |
+| `$HOME/.claude/skills/xcode-project-analyzer/SKILL.md` | Build settings / script phases / parallelism |
+| `$HOME/.claude/skills/spm-build-analysis/SKILL.md` | SPM graph + plugins |
+| `$HOME/.claude/skills/xcode-build-fixer/SKILL.md` | Apply approved fixes + re-benchmark |
+| `$HOME/.claude/skills/NOTICE-xcode-build-skills.md` | MIT attribution + upstream pin |
 
 ## Notes
 

package/pipeline/commands/multi-agent/channels/SKILL.md

@@ -73,7 +73,7 @@ Rendered in English (`promptLanguage` is locked to `en`). Previous selections (`
 **Runs only when** (a) `agent-state.json` exists for the task AND (b) a Work Summary can be generated (same availability rule as the "Yapılan iş özeti" content row). Purpose: give the user an instant "what did this ship?" answer BEFORE they pick channels. They see the outcome first, then decide where to route it  -  instead of the reverse.
 
 ```bash
-preview=$(bash pipeline/scripts/render-work-summary.sh "$TASK_ID" 2>/dev/null || true)
+preview=$(bash $HOME/.claude/scripts/render-work-summary.sh "$TASK_ID" 2>/dev/null || true)
 [ -n "$preview" ] && {
   printf '%s\n\n' "$preview"
   # Auto-tick workSummary content row when preview was shown.
@@ -240,7 +240,7 @@ If no cached report exists, skip silently  -  this is augmentation, not a gate.
 
 **Work summary generation:**
 
-Emitted only when `reportContent.workSummary === true` and at least one of (`agent-state.json`, `--branch` flag) is available. The shell adapter is `pipeline/scripts/render-work-summary.sh <taskId>`:
+Emitted only when `reportContent.workSummary === true` and at least one of (`agent-state.json`, `--branch` flag) is available. The shell adapter is `$HOME/.claude/scripts/render-work-summary.sh <taskId>`:
 
 1. **Task header**  -  `taskId`, `branch`, `baseBranch`, `prNumber` from `agent-state.json` (or explicit flags for post-hoc invocation).
 2. **Scope delivered**  -  Phase 2 `planTodos[]` / `tasks[]` rendered as `✅` (status=done) or `⏳` (anything else) rows. Task id + title shown; `(deferred  -  rationale)` appended if the task's `status` is `"deferred"`.
@@ -278,7 +278,7 @@ Emitted only when `reportContent.costSummary === true` and tracker data is avail
 
 1. **Token tallies per phase**  -  read `.worktrees/{taskPath}/phase-tracker.json` (written by `phase-tracker.sh update` / `tokens` actions). If present, each phase row has `{tokens_in, tokens_out}`.
 2. **Fallback to OTel spans**  -  if tracker JSON has no token columns but `otel-spans.jsonl` exists (set `MULTI_AGENT_OTEL_SPANS=1` beforehand), aggregate `phase.tokens` spans by phase ID.
-3. **USD estimate**  -  uses a static price table in `pipeline/scripts/cost-table.json` (model → $/Mtok_in, $/Mtok_out). Missing model → show ` - ` for USD, never block.
+3. **USD estimate**  -  uses a static price table in `$HOME/.claude/scripts/cost-table.json` (model → $/Mtok_in, $/Mtok_out). Missing model → show ` - ` for USD, never block.
 
 **Output template:**
 
@@ -293,7 +293,7 @@ Emitted only when `reportContent.costSummary === true` and tracker data is avail
 | **Total**   |        | **94,744** | **22,930** | **$2.04** |
 ```
 
-Generated by `pipeline/scripts/render-cost-summary.sh <task-id>` using prices from `pipeline/scripts/cost-table.json`. The channels adapter shells out to this renderer  -  doc table is illustrative only, runtime is authoritative.
+Generated by `$HOME/.claude/scripts/render-cost-summary.sh <task-id>` using prices from `$HOME/.claude/scripts/cost-table.json`. The channels adapter shells out to this renderer  -  doc table is illustrative only, runtime is authoritative.
 
 **Precision rules:** tokens as thousand-separated integers. USD rounded to 2 dp (`printf '$%.2f'`). Rows with zero tokens dropped. If total USD is ` - ` (no price for any model), replace the total row USD cell with ` - ` and add a one-line footnote: `*USD unavailable  -  add model to cost-table.json*`.
 
@@ -469,7 +469,7 @@ Phase 7 passes a runtime state bundle: `{jiraId, prUrl, remoteType, taskType, fi
 | Claude Code | `/multi-agent:channels` | Slash namespace (colon) |
 | Copilot CLI | `multi-agent-channels` | Top-level command (dash-separated, no namespace support) |
 
-Both installations read the same source file (`pipeline/commands/multi-agent/channels/SKILL.md`)  -  `install.js --all` copies to:
+Both installations read the same source file (`$HOME/.claude/commands/multi-agent/channels/SKILL.md`)  -  `install.js --all` copies to:
 - `~/.claude/commands/multi-agent/channels/SKILL.md`
 - `~/.copilot/commands/multi-agent-channels.md`
 

package/pipeline/commands/multi-agent/dev/SKILL.md

@@ -1,6 +1,6 @@
 ---
-description: "Fast development mode: Init → Dev (Opus) → Test → Commit → Report. Analysis, planning, and review phases are skipped. Use when the work is already scoped and only development, test and commit are needed."
-description-tr: "Hızlı geliştirme modu: Init → Dev (Opus) → Test → Commit → Report. Analiz, planlama ve review fazları atlanır."
+description: "Fast development mode: Init → Dev (Opus) → Review → Test → Commit → Report. Analysis and planning are skipped, review is not. Use when the work is already scoped but the result still has to be checked."
+description-tr: "Hızlı geliştirme modu: Init → Dev (Opus) → Review → Test → Commit → Report. Analiz ve planlama atlanır, review atlanmaz."
 ---
 
 # multi-agent dev  -  Fast Development Mode (--dev)
@@ -21,28 +21,31 @@ description-tr: "Hızlı geliştirme modu: Init → Dev (Opus) → Test → Comm
 >
 > Full contract: `$HOME/.claude/multi-agent-refs/rules.md` "Language Application".
 
-A 5-phase fast pipeline: Init → Dev → Test → Commit → Report. Development runs on the Opus model (top intelligence tier).
+A 6-phase fast pipeline: Init → Dev → Review → Test → Commit → Report. Development runs on the Opus model (top intelligence tier).
 
 ## When to use it
 - Small changes, bug fixes, quick features
 - Simple tasks that don't need the full 8-phase pipeline
-- Routine work that does not require formal review
+- Work that is already scoped, so analysis and planning buy nothing  -  the result is still reviewed
 
 ## Pipeline
 
 ```
 Phase 0: Init       → project detection, worktree, branch, identity (full picker)
 Phase 3: Dev        → direct development on Opus (TDD optional)
+Phase 4: Review     → deterministic gates + parallel review + triage
 Phase 5: Test       → User Test (interactive simulator)
 Phase 6: Commit     → commit + push + PR
 Phase 7: Report     → channels (Jira / Confluence / PR / Wiki)
 ```
 
-`--dev` skips Phase 1 (Analysis), Phase 2 (Planning + Approval Gate), and Phase 4 (Review). Phase 3 Dev runs on **Opus**. Phase 0 picker, Phase 5 test, Phase 6 commit, and Phase 7 channels are identical to the full pipeline.
+`--dev` skips Phase 1 (Analysis) and Phase 2 (Planning + Approval Gate). Phase 3 Dev runs on **Opus**. Phase 0 picker, Phase 4 review, Phase 5 test, Phase 6 commit, and Phase 7 channels are identical to the full pipeline.
 
 ## What `--dev` does NOT skip (required)
 
-`--dev` skips the LLM-heavy phases (Phase 1 Analysis, Phase 2 Planning + Approval, Phase 4 Review) and runs Phase 3 Dev on Opus. Phase 0 (Init), Phase 5 (Test), Phase 6 (Commit), and Phase 7 (Report + channels) are byte-for-byte the same as the full pipeline. User-facing prompts **always run**  -  they exist for audit, not speed.
+`--dev` skips the two phases that front-load a task (Phase 1 Analysis, Phase 2 Planning + Approval) and runs Phase 3 Dev on Opus. Phase 0 (Init), Phase 4 (Review), Phase 5 (Test), Phase 6 (Commit), and Phase 7 (Report + channels) are byte-for-byte the same as the full pipeline. User-facing prompts **always run**  -  they exist for audit, not speed.
+
+**Review is not a fast-mode casualty.** Skipping analysis means the run has less context, not that the output deserves less scrutiny: Phase 4 runs its deterministic gates, dispatches the parallel reviewers and triages their findings exactly as the full pipeline does. Accepted blocking findings loop back to Phase 3, capped at 3 iterations. Because Phase 1 never ran, Phase 4 substitutes a deterministic stack classifier for `detectedStack` and records every Phase-1-dependent step as `not-applicable`, so a step that could not run is distinguishable from a step that passed.
 
 ### Required user prompts (Phase 0)
 
@@ -155,9 +158,8 @@ No "pre-existing" claim is allowed without a baseline reproduce.
 
 - Phase 1 (Analysis → Opus deep-think + explore agents)
 - Phase 2 → Plan Approval Gate including clarification + approval
-- Phase 4 (Review → parallel + triage)
 
-Phase 0 (Init full picker), Phase 5 (User Test), Phase 6 (Commit), and Phase 7 (Report + channels) run exactly as in the full pipeline. Phase 3 Dev model is **Opus**.
+That is the whole list. Phase 0 (Init full picker), Phase 4 (Review), Phase 5 (User Test), Phase 6 (Commit), and Phase 7 (Report + channels) run exactly as in the full pipeline. Phase 3 Dev model is **Opus**.
 
 ## Steps
 
@@ -187,25 +189,28 @@ Phase 0 (Init full picker), Phase 5 (User Test), Phase 6 (Commit), and Phase 7 (
 
    The dispatch is keyed on `agent-state.taskType`, written by Phase 0 Step 7 (`feedback_phase0_tasktype.md`). If the issue body contains a Figma URL, the type is `"component"` and 3a applies  -  no exception, even in `--dev`.
 
-4. **Phase 5: User Test**  -  same as the full pipeline (interactive simulator / local-test prompt)
+4. **Phase 4: Review**  -  same as the full pipeline, per `$HOME/.claude/multi-agent-refs/phases/phase-4-review.md`: deterministic gates (build / lint / test / secrets / evidence), the parallel reviewer set (CLI-aware), and triage. Accepted blocking findings return to step 3 for rework, capped at 3 iterations per `operations.md`. Phase-1-dependent steps (1.8 Figma evidence, 2.8 visual conformance) are recorded as `not-applicable (no Phase 1 evidence)` rather than skipped silently.
+
+5. **Phase 5: User Test**  -  same as the full pipeline (interactive simulator / local-test prompt)
 
-5. **Phase 6: Commit**  -  same as the full pipeline (commit + push + PR + shared-branch final confirm)
+6. **Phase 6: Commit**  -  same as the full pipeline (commit + push + PR + shared-branch final confirm). Phase 6 refuses to commit while an accepted blocking finding is unresolved.
 
-6. **Phase 7: Report**  -  same as the full pipeline (channels prompt: Jira / Confluence / PR description / Wiki)
+7. **Phase 7: Report**  -  same as the full pipeline (channels prompt: Jira / Confluence / PR description / Wiki)
 
 ## Differences vs full pipeline
 
 | Aspect | Full (`/multi-agent`) | Dev (`/multi-agent:dev`) |
 |---------|---------------------|------------------------|
-| Phases | 8 (0-7) | 5 (0, 3, 5, 6, 7)  -  1, 2, 4 skipped |
+| Phases | 8 (0-7) | 6 (0, 3, 4, 5, 6, 7)  -  1, 2 skipped |
 | Phase 3 Dev model | Sonnet | **Opus** |
 | Phase 1 Analysis (explore agents) | ✅ | ❌ |
 | Phase 2 Planning + Approval Gate | ✅ | ❌ |
-| Phase 4 Review (parallel + triage) | ✅ | ❌ |
+| Phase 4 Review (parallel + triage) | ✅ | ✅ (same) |
+| Phase 4 criteria source | Phase 1 `detectedStack` | deterministic stack classifier |
 | Phase 0 picker (account, repos, branch, ...) | ✅ | ✅ (same) |
 | Phase 5 User Test | ✅ | ✅ (same) |
 | Phase 7 channels (Jira / Confluence / PR / Wiki) | ✅ | ✅ (same) |
-| Duration | ~10-15 min | ~5-7 min |
+| Duration | ~10-15 min | ~7-10 min |
 ## Required: Phase Tracker Contract
 
 **The phase tracker is mandatory**  -  the agent cannot skip it. Full spec: [`$HOME/.claude/multi-agent-refs/tracker-contract.md`]($HOME/.claude/multi-agent-refs/tracker-contract.md).
@@ -218,7 +223,7 @@ Two channels run in parallel at every phase boundary:
 ```bash
 # Phase 0, very first shell call (every CLI):
 bash $HOME/.claude/scripts/phase-tracker.sh init "$TASK_ID"
-for p in "0:Init" "3:Dev" "5:Test" "6:Commit" "7:Report"; do
+for p in "0:Init" "3:Dev" "4:Review" "5:Test" "6:Commit" "7:Report"; do
   bash $HOME/.claude/scripts/phase-tracker.sh add "${p%%:*}" "${p#*:}"
 done
 bash $HOME/.claude/scripts/phase-tracker.sh update 0 in_progress
@@ -238,7 +243,7 @@ In Claude Code the agent MUST also drive the native TaskList widget so the user
 
 ```text
 # Phase 0 startup  -  register one tile per phase (0..N), capture the taskId, persist it:
-for each phase in 0:Init, 3:Dev, 5:Test, 6:Commit, 7:Report:
+for each phase in 0:Init, 3:Dev, 4:Review, 5:Test, 6:Commit, 7:Report:
   TaskCreate({ subject: "Phase <N>: <Name>", activeForm: "<doing-form>" })
     -> returns taskId
   bash $HOME/.claude/scripts/phase-tracker.sh meta <N> tasklist_id "<taskId>"
@@ -255,11 +260,11 @@ TaskUpdate({ taskId: <saved>, status: "completed" })
 bash $HOME/.claude/scripts/phase-tracker.sh update <N> completed
 ```
 
-`--dev` mode does NOT TaskCreate phases 1/2/4  -  those are not part of the `--dev` phase set (`0:Init 3:Dev 5:Test 6:Commit 7:Report`). Only register tiles for the active set.
+`--dev` mode does NOT TaskCreate phases 1/2  -  those are not part of the `--dev` phase set (`0:Init 3:Dev 4:Review 5:Test 6:Commit 7:Report`). Only register tiles for the active set.
 
 #### TaskCreate ordering (strict)
 
-**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev` that means: Phase 0 → Phase 3 → Phase 5 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
+**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev` that means: Phase 0 → Phase 3 → Phase 4 → Phase 5 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
 
 ### Visual channel  -  Copilot CLI / plain shell
 

package/pipeline/commands/multi-agent/dev-autopilot/SKILL.md

@@ -1,6 +1,6 @@
 ---
-description: "Fastest mode: Dev (Opus) plus Autopilot. Init → Dev → Commit → Report with zero confirmations. Use when a change is well understood and should go from start to commit with no questions asked."
-description-tr: "En hızlı mod: Dev (Opus) + Autopilot. Init → Dev → Commit → Report, sıfır onay."
+description: "Fastest mode: Dev (Opus) plus Autopilot. Init → Dev → Review → Commit → Report with zero confirmations. Review still runs and auto-fixes blocking findings. Use when a change is well understood and should go from start to commit with no questions asked."
+description-tr: "En hızlı mod: Dev (Opus) + Autopilot. Init → Dev → Review → Commit → Report, sıfır onay. Review koşar, blocking bulguları otomatik düzeltir."
 ---
 
 # multi-agent dev autopilot  -  Fastest Path
@@ -21,38 +21,44 @@ Dev mode + Autopilot combined: a 4-phase pipeline with no confirmations, end-to-
 ```
 Phase 0: Init       → project detection, worktree, branch, identity
 Phase 3: Dev        → direct development on Opus
+Phase 4: Review     → gates + parallel review + triage, auto-fix
 Phase 6: Commit     → auto commit + push + PR (no confirmations)
 Phase 7: Report     → short terminal summary
 ```
 
-**Skipped phases**: Analysis (1), Planning (2), Review (4), User Test (5).
+**Skipped phases**: Analysis (1), Planning (2), User Test (5).
 **Skipped confirmations**: Plan Approval Gate (clarification + approval loop), test prompt, commit prompt, PR prompt.
 
 > If you want to talk through the plan, do not use this mode  -  pick `/multi-agent "task"` (normal) or any mode without `autopilot`. The Dev+Autopilot contract is: "ask nothing."
 
+**Review runs here too, and asking nothing does not mean accepting anything.** Autopilot consumes the *confirmation* prompts, not the *quality* gates: Phase 4 reviews the diff and auto-fixes accepted blocking findings without asking. The contract holds because no question is put to the user  -  the run only halts when it cannot converge, which is the rework-storm trigger in `$HOME/.claude/multi-agent-refs/features/autopilot-circuit-breaker.md` (`maxReworkCycles = 3`).
+
 ## Steps
 
 1. **Parse input**  -  standard multi-agent input formats (Issue URL, Jira ID, free text)
 2. **Phase 0: Init**  -  set `"mode": "dev", "autopilot": true` in `agent-state.json`
-3. **Phase 3: Dev**  -  write code directly on `claude-opus-4-8` and verify the build
-4. **Phase 6: Commit**  -  auto commit + push + PR
-5. **Phase 7: Report**  -  terminal summary
+3. **Phase 3: Dev**  -  write code directly on `claude-opus-5` and verify the build
+4. **Phase 4: Review**  -  gates, parallel reviewers, triage. Accepted blocking findings are fixed automatically (back to step 3), no prompt
+5. **Phase 6: Commit**  -  auto commit + push + PR
+6. **Phase 7: Report**  -  terminal summary
 
 ## Safety
 
 - Build failure → 3 retries; if it still fails → **pause** (ask the user)
+- Review rework → 3 cycles; if blocking findings survive → **pause** (circuit breaker, no auto-commit)
 - Kill / Purge → always asks for confirmation (destructive)
 
 ## Differences
 
 | | Full | Dev | Autopilot | **Dev+Autopilot** |
 |--|------|-----|-----------|-------------------|
-| Phases | 8 | 4 | 8 | **4** |
+| Phases | 8 | 6 | 8 | **5** |
 | Model | Sonnet | Opus | Sonnet | **Opus** |
 | Plan Approval Gate | ✅ (clarification + approval) | ❌ | ❌ | **❌** |
 | Confirmations (test / commit / PR) | Yes | Yes | No | **No** |
-| Review | Parallel + triage (CLI-aware) | None | Parallel + triage (CLI-aware) | **None** |
-| Estimated duration | ~12 min | ~5 min | ~10 min | **~3 min** |
+| Review | Parallel + triage (CLI-aware) | Parallel + triage (CLI-aware) | Parallel + triage (CLI-aware) | **Parallel + triage, auto-fix** |
+| Blocking finding | Fix loop, then ask | Fix loop, then ask | Auto-fix, breaker at 3 | **Auto-fix, breaker at 3** |
+| Estimated duration | ~12 min | ~7 min | ~10 min | **~5 min** |
 ## Required: Phase Tracker Contract
 
 **The phase tracker is mandatory**  -  the agent cannot skip it. Full spec: [`$HOME/.claude/multi-agent-refs/tracker-contract.md`]($HOME/.claude/multi-agent-refs/tracker-contract.md).
@@ -67,7 +73,7 @@ Two channels run in parallel at every phase boundary:
 ```bash
 # Phase 0, very first shell call (every CLI):
 bash $HOME/.claude/scripts/phase-tracker.sh init "$TASK_ID"
-for p in "0:Init" "3:Dev" "6:Commit" "7:Report"; do
+for p in "0:Init" "3:Dev" "4:Review" "6:Commit" "7:Report"; do
   bash $HOME/.claude/scripts/phase-tracker.sh add "${p%%:*}" "${p#*:}"
 done
 bash $HOME/.claude/scripts/phase-tracker.sh update 0 in_progress
@@ -87,7 +93,7 @@ In Claude Code the agent MUST also drive the native TaskList widget so the user
 
 ```text
 # Phase 0 startup  -  register one tile per phase (0..N), capture the taskId, persist it:
-for each phase in 0:Init, 3:Dev, 6:Commit, 7:Report:
+for each phase in 0:Init, 3:Dev, 4:Review, 6:Commit, 7:Report:
   TaskCreate({ subject: "Phase <N>: <Name>", activeForm: "<doing-form>" })
     -> returns taskId
   bash $HOME/.claude/scripts/phase-tracker.sh meta <N> tasklist_id "<taskId>"
@@ -104,11 +110,11 @@ TaskUpdate({ taskId: <saved>, status: "completed" })
 bash $HOME/.claude/scripts/phase-tracker.sh update <N> completed
 ```
 
-`--dev autopilot` mode does NOT TaskCreate phases 1/2/4/5  -  those are not part of the `--dev autopilot` phase set (`0:Init 3:Dev 6:Commit 7:Report`). Only register tiles for the active set.
+`--dev autopilot` mode does NOT TaskCreate phases 1/2/5  -  those are not part of the `--dev autopilot` phase set (`0:Init 3:Dev 4:Review 6:Commit 7:Report`). Only register tiles for the active set.
 
 #### TaskCreate ordering (strict)
 
-**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev autopilot` that means: Phase 0 → Phase 3 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
+**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev autopilot` that means: Phase 0 → Phase 3 → Phase 4 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
 
 ### Visual channel  -  Copilot CLI / plain shell
 

package/pipeline/commands/multi-agent/dev-local/SKILL.md

@@ -1,6 +1,6 @@
 ---
-description: "Fast mode + local  -  Init → Dev(Opus) → Commit → Report, no worktree. Use when a change should be developed on the current branch without creating a worktree."
-description-tr: "Hızlı mod + lokal - Init → Dev(Opus) → Commit → Report, worktree yok."
+description: "Fast mode + local  -  Init → Dev(Opus) → Review → Commit → Report, no worktree. Use when a change should be developed and reviewed on the current branch without creating a worktree."
+description-tr: "Hızlı mod + lokal - Init → Dev(Opus) → Review → Commit → Report, worktree yok."
 allowed-tools: Agent, Bash, Read, Write, Edit, Glob, Grep, TaskCreate, TaskUpdate, TaskList, TaskGet, AskUserQuestion, WebFetch, Skill
 ---
 
@@ -14,20 +14,21 @@ Dedicated form of the `--dev` + `--local` combination. A 4-phase fast pipeline (
 
 ```
 Phase 0: Init       → project detection, branch check, state (NO worktree)
-Phase 3: Dev        → direct development on Opus (Analysis + Planning + Review skipped)
+Phase 3: Dev        → direct development on Opus (Analysis + Planning skipped)
+Phase 4: Review     → deterministic gates + parallel review + triage
 Phase 6: Commit     → pre-commit checkout prompt, commit + push + PR
 Phase 7: Report     → Jira / Wiki + log + knowledge/memory
 ```
 
-`--dev local` skips Phase 1 (Analysis), Phase 2 (Planning + Approval Gate), Phase 4 (Review), and Phase 5 (User Test  -  local/autopilot variants skip the interactive test gate). It differs from `--dev` on two axes: no git worktree is created (development happens directly on the current branch in `$PROJECT_ROOT`), and the interactive User Test phase is skipped.
+`--dev local` skips Phase 1 (Analysis), Phase 2 (Planning + Approval Gate), and Phase 5 (User Test  -  local/autopilot variants skip the interactive test gate). It differs from `--dev` on two axes: no git worktree is created (development happens directly on the current branch in `$PROJECT_ROOT`), and the interactive User Test phase is skipped. **Review is not skipped**  -  Phase 4 runs its gates, resolves the criteria the dev phase was supposed to honour, reviews in parallel and triages, and accepted blocking findings return to Phase 3 (3-iteration cap).
 
-> **Want the quality tail afterwards?** Since this mode skips Review + Test, run [`/multi-agent:finish`](../finish/SKILL.md) on the same branch when you're done to add parallel review, a build+test success gate, PR, and a Jira technical-analysis + test-scenario comment  -  without re-developing.
+> **Already have work on the branch?** For changes developed outside the pipeline, or by hand, run [`/multi-agent:ship`](../ship/SKILL.md) on that branch: it puts the existing diff through the same review, adds a build+test success gate, opens the PR and posts the Jira technical-analysis + test-scenario comment  -  without re-developing.
 
 ## When to use it
 
-- Quick bug fix or small feature  -  review overhead is unnecessary
+- Quick bug fix or small feature  -  analysis and planning buy nothing here
 - You're working on the current branch and don't want to switch worktrees
-- A change you trust  -  no need for multi-model review
+- A well-understood change  -  you want it reviewed, not designed
 ## Required: Phase Tracker Contract
 
 **The phase tracker is mandatory**  -  the agent cannot skip it. Full spec: [`$HOME/.claude/multi-agent-refs/tracker-contract.md`]($HOME/.claude/multi-agent-refs/tracker-contract.md).
@@ -42,7 +43,7 @@ Two channels run in parallel at every phase boundary:
 ```bash
 # Phase 0, very first shell call (every CLI):
 bash $HOME/.claude/scripts/phase-tracker.sh init "$TASK_ID"
-for p in "0:Init" "3:Dev" "6:Commit" "7:Report"; do
+for p in "0:Init" "3:Dev" "4:Review" "6:Commit" "7:Report"; do
   bash $HOME/.claude/scripts/phase-tracker.sh add "${p%%:*}" "${p#*:}"
 done
 bash $HOME/.claude/scripts/phase-tracker.sh update 0 in_progress
@@ -62,7 +63,7 @@ In Claude Code the agent MUST also drive the native TaskList widget so the user
 
 ```text
 # Phase 0 startup  -  register one tile per phase (0..N), capture the taskId, persist it:
-for each phase in 0:Init, 3:Dev, 6:Commit, 7:Report:
+for each phase in 0:Init, 3:Dev, 4:Review, 6:Commit, 7:Report:
   TaskCreate({ subject: "Phase <N>: <Name>", activeForm: "<doing-form>" })
     -> returns taskId
   bash $HOME/.claude/scripts/phase-tracker.sh meta <N> tasklist_id "<taskId>"
@@ -79,11 +80,11 @@ TaskUpdate({ taskId: <saved>, status: "completed" })
 bash $HOME/.claude/scripts/phase-tracker.sh update <N> completed
 ```
 
-`--dev local` mode does NOT TaskCreate phases 1/2/4/5  -  those are not part of the `--dev local` phase set (`0:Init 3:Dev 6:Commit 7:Report`). Only register tiles for the active set.
+`--dev local` mode does NOT TaskCreate phases 1/2/5  -  those are not part of the `--dev local` phase set (`0:Init 3:Dev 4:Review 6:Commit 7:Report`). Only register tiles for the active set.
 
 #### TaskCreate ordering (strict)
 
-**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev local` that means: Phase 0 → Phase 3 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
+**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev local` that means: Phase 0 → Phase 3 → Phase 4 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
 
 ### Visual channel  -  Copilot CLI / plain shell
 
@@ -98,8 +99,9 @@ Do NOT call TaskCreate on these CLIs  -  the tool does not exist and the call fa
 
 Routes to the orchestrator with `--dev --local` flags. Apply the `$HOME/.claude/commands/multi-agent/dev/SKILL.md` pipeline in local mode:
 - Phase 0: skip worktree creation, continue on the current branch
-- Phase 1 (Analysis), Phase 2 (Planning + Approval Gate), Phase 4 (Review), and Phase 5 (User Test) are skipped (`--dev` + local/autopilot drop the interactive test gate)
+- Phase 1 (Analysis), Phase 2 (Planning + Approval Gate), and Phase 5 (User Test) are skipped (`--dev` + local/autopilot drop the interactive test gate)
 - Phase 3: develop on Opus (instead of the Sonnet TDD cycle)
+- Phase 4: review as in the full pipeline, on the local branch diff; accepted blocking findings loop back to Phase 3, capped at 3 iterations
 - Phase 6: commit + push + PR (the local checkout prompt is natural in local mode  -  you're already there; no worktree removal needed, code is already in `$PROJECT_ROOT`)
 - Phase 7: report + channels (same as `--dev`)
 

package/pipeline/commands/multi-agent/dev-local-autopilot/SKILL.md

@@ -1,6 +1,6 @@
 ---
-description: "Fastest + local  -  Dev(Opus) + autopilot, no worktree, zero interaction. Use when a change should be developed on the current branch with no worktree and no prompts."
-description-tr: "En hızlı + lokal - Dev(Opus) + autopilot, worktree yok, sıfır etkileşim."
+description: "Fastest + local  -  Dev(Opus) + Review + autopilot, no worktree, zero interaction. Review runs and auto-fixes blocking findings. Use when a change should be developed on the current branch with no worktree and no prompts."
+description-tr: "En hızlı + lokal - Dev(Opus) + Review + autopilot, worktree yok, sıfır etkileşim. Review koşar, blocking bulguları otomatik düzeltir."
 allowed-tools: Agent, Bash, Read, Write, Edit, Glob, Grep, TaskCreate, TaskUpdate, TaskList, TaskGet, WebFetch, Skill
 ---
 
@@ -15,25 +15,30 @@ The triple `--dev` + `--local` + `autopilot`  -  the fastest form available. Zer
 ```
 Phase 0: Init       → project detection, branch check, state (NO worktree, NO confirmation)
 Phase 3: Dev        → direct development on Opus (automatic)
+Phase 4: Review     → gates + parallel review + triage, auto-fix
 Phase 6: Commit     → auto commit + push + PR (no local checkout prompt)
 Phase 7: Report     → Jira / Wiki + log + knowledge/memory
 ```
 
+Phase 1 (Analysis), Phase 2 (Planning + Approval Gate) and Phase 5 (User Test) are skipped. Review is not: accepted blocking findings are fixed automatically without a prompt, and the run halts rather than committing if they survive 3 rework cycles.
+
 ## When to use it
 
-- Routine change you trust that needs no review
+- Routine change you trust  -  you want it reviewed, not discussed
 - A CI / batch environment running the pipeline automatically  -  no user interaction
-- Prototype / spike  -  you care about speed, quality gates can wait
+- Prototype / spike  -  you care about speed, and the review that runs costs you no prompts
 
 ## When NOT to use it
 
-- Production-bound change  -  at minimum use `--dev` (without autopilot)
-- Security-sensitive code  -  autopilot bypasses the safety gates
-- A first big refactor  -  human review is required
+- Production-bound change  -  at minimum use `--dev` (without autopilot), so you see the findings before the commit
+- Security-sensitive code  -  autopilot decides for you which accepted findings to fix, and commits when they are gone
+- A first big refactor  -  human review is required, and machine review does not substitute for it
 
 ## Delegation
 
-Routes to the orchestrator with `--dev --local autopilot` flags. The pipeline contract matches `dev-autopilot.md` exactly, with Phase 0 Step 8 (worktree creation) skipped.
+Routes to the orchestrator with `--dev --local autopilot` flags. The pipeline contract matches [`dev-autopilot/SKILL.md`](../dev-autopilot/SKILL.md) exactly, with Phase 0 Step 8 (worktree creation) skipped.
+
+> **Already have work on the branch?** For a diff that exists without a pipeline run behind it, [`/multi-agent:ship`](../ship/SKILL.md) applies the same review plus a build+test gate before the PR.
 
 ## Examples
 
@@ -57,7 +62,7 @@ Two channels run in parallel at every phase boundary:
 ```bash
 # Phase 0, very first shell call (every CLI):
 bash $HOME/.claude/scripts/phase-tracker.sh init "$TASK_ID"
-for p in "0:Init" "3:Dev" "6:Commit" "7:Report"; do
+for p in "0:Init" "3:Dev" "4:Review" "6:Commit" "7:Report"; do
   bash $HOME/.claude/scripts/phase-tracker.sh add "${p%%:*}" "${p#*:}"
 done
 bash $HOME/.claude/scripts/phase-tracker.sh update 0 in_progress
@@ -77,7 +82,7 @@ In Claude Code the agent MUST also drive the native TaskList widget so the user
 
 ```text
 # Phase 0 startup  -  register one tile per phase (0..N), capture the taskId, persist it:
-for each phase in 0:Init, 3:Dev, 6:Commit, 7:Report:
+for each phase in 0:Init, 3:Dev, 4:Review, 6:Commit, 7:Report:
   TaskCreate({ subject: "Phase <N>: <Name>", activeForm: "<doing-form>" })
     -> returns taskId
   bash $HOME/.claude/scripts/phase-tracker.sh meta <N> tasklist_id "<taskId>"
@@ -94,11 +99,11 @@ TaskUpdate({ taskId: <saved>, status: "completed" })
 bash $HOME/.claude/scripts/phase-tracker.sh update <N> completed
 ```
 
-`--dev local autopilot` mode does NOT TaskCreate phases 1/2/4/5  -  those are not part of the `--dev local autopilot` phase set (`0:Init 3:Dev 6:Commit 7:Report`). Only register tiles for the active set.
+`--dev local autopilot` mode does NOT TaskCreate phases 1/2/5  -  those are not part of the `--dev local autopilot` phase set (`0:Init 3:Dev 4:Review 6:Commit 7:Report`). Only register tiles for the active set.
 
 #### TaskCreate ordering (strict)
 
-**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev local autopilot` that means: Phase 0 → Phase 3 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
+**All TaskCreate calls fire in strict phase-number order BEFORE any TaskUpdate is applied.** For `--dev local autopilot` that means: Phase 0 → Phase 3 → Phase 4 → Phase 6 → Phase 7. The native widget renders by creation order, not by phase number  -  out-of-order calls produce visually scrambled tile stacks. Full ordering contract in `$HOME/.claude/multi-agent-refs/tracker-contract.md` section "TaskCreate ordering (strict)".
 
 ### Visual channel  -  Copilot CLI / plain shell
 

package/pipeline/commands/multi-agent/garbage-collect/SKILL.md

@@ -62,7 +62,7 @@ deletes nothing until you confirm.
    residue in the index (`.worktrees/{id}` recorded as a "Subproject commit"
    entry by a pre-guard `git add -A`), and a missing `.worktrees/` line in
    `.git/info/exclude`. Registered, healthy worktrees are NEVER touched  -
-   those belong to `/multi-agent:finish` / `/multi-agent:kill`.
+   those belong to `/multi-agent:ship` / `/multi-agent:kill`.
 
    - Output says `nothing to do` -> skip silently, no question.
    - Otherwise surface a second `AskUserQuestion` (in `outputLanguage`):