maestro-flow 0.5.37 → 0.5.39

package/workflows/ui-design.md

@@ -124,7 +124,7 @@ python3 "${SKILL_PATH}" "accessibility animation interaction" --domain ux
 #### 3b. Generate design direction options via ui-design-agent
 
 ```
-Agent(ui-design-agent): [DESIGN_DIRECTION_GENERATION]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [DESIGN_DIRECTION_GENERATION]
 
 Generate ${styleCount} maximally contrasting design directions (min 6D distance: 0.7).
 
@@ -152,7 +152,7 @@ Update `analysis-options.json` with `user_selection` field.
 For each selected direction, spawn ui-design-agent:
 
 ```
-Agent(ui-design-agent): [DESIGN_SYSTEM_GENERATION #${variant_index}]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [DESIGN_SYSTEM_GENERATION #${variant_index}]
 
 Generate production-ready design tokens from selected direction.
 Map 6D attributes to token values:
@@ -200,7 +200,7 @@ Requirements: ALL colors OKLCH, WCAG AA (4.5:1 text, 3:1 UI), complete combinati
 **Purpose:** Generate animation system complementing the selected styles.
 
 ```
-Agent(ui-design-agent): [ANIMATION_SYSTEM_GENERATION]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [ANIMATION_SYSTEM_GENERATION]
 
 Input: selected variant's design-tokens.json (mood/weight context) + ui-ux-pro-max UX data.
 Output: ${PHASE_DIR}/design-ref/animation-tokens.json
@@ -227,7 +227,7 @@ Requirements: CSS custom properties, prefers-reduced-motion mandatory, var() ref
 #### 5a. Generate layout concept options
 
 ```
-Agent(ui-design-agent): [LAYOUT_CONCEPT_GENERATION]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [LAYOUT_CONCEPT_GENERATION]
 
 Generate ${layoutCount} structurally distinct layout concepts per target.
 Concepts must differ in: grid structure, component arrangement, visual hierarchy, navigation pattern.
@@ -249,7 +249,7 @@ Update analysis-options.json with `user_selection.selected_variants`.
 For each selected concept × target:
 
 ```
-Agent(ui-design-agent): [LAYOUT_TEMPLATE_GENERATION -- ${target} variant ${variant_id}]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [LAYOUT_TEMPLATE_GENERATION -- ${target} variant ${variant_id}]
 
 Structure ONLY, no visual style. Output: ${PHASE_DIR}/design-ref/layout-templates/layout-${target}-${variant_id}.json
 
@@ -267,6 +267,8 @@ Schema:
 Rules: semantic HTML5, ARIA roles, var() refs only (no hard-coded values), mobile-first responsive. NO colors/fonts/shadows.
 ```
 
+GATE Step 5→6: design-tokens.json + animation-tokens.json + layout-templates exist BEFORE assembly; BLOCKED if any missing
+
 ---
 
 ### Step 6: Assemble HTML Prototypes (Layer 4 — Assembly)
@@ -276,7 +278,7 @@ Rules: semantic HTML5, ARIA roles, var() refs only (no hard-coded values), mobil
 For each `style × layout × target` combination, spawn ui-design-agent:
 
 ```
-Agent(ui-design-agent): [PROTOTYPE_ASSEMBLY -- ${target}-style-${s}-layout-${l}]
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Agent(ui-design-agent): [PROTOTYPE_ASSEMBLY -- ${target}-style-${s}-layout-${l}]
 
 Pure assembly: combine pre-extracted structure + tokens. NO design decisions.
 
@@ -375,7 +377,7 @@ Spacing & Layout, Effects & Interactions, Component Styles, Animation System, An
 | Python not found | Abort with install instructions per OS |
 | ui-ux-pro-max not found | Abort, suggest skill installation |
 | Design system returns empty | Retry with broader keywords, then abort |
-| Prototype agent fails | Log error, continue with other variants |
+| Prototype agent fails | Log error, continue with other variants; flag variants as [LOW CONFIDENCE] (prototype agent failed) |
 | User cancels selection | Save all variants as-is, exit without MASTER.md |
 | --refine without existing design-ref | Error E004 |
 

package/workflows/ui-style.md

@@ -94,6 +94,8 @@ Save variants to `${PHASE_DIR}/design-ref/prototypes/variant-{N}-system.md`.
 Display each variant summary (pattern, colors, typography, effects, anti-patterns).
 User selects: [1-N | "redo" | "all"]. Auto mode selects variant 1.
 
+GATE Step 3→4: selected variant confirmed BEFORE solidification; BLOCKED if no variant selected
+
 ---
 
 ### Step 4: Solidify Selected Design
@@ -109,7 +111,7 @@ python3 "${SKILL_PATH}" "${selected_variant_keywords}" --design-system --persist
 
 #### 4b. Generate design-tokens.json
 
-Spawn agent to extract structured tokens from MASTER.md into `${PHASE_DIR}/design-ref/design-tokens.json`.
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Spawn agent to extract structured tokens from MASTER.md into `${PHASE_DIR}/design-ref/design-tokens.json`.
 
 **Token schema keys:** colors (brand/surface/semantic/text/border in OKLCH), typography (family/size/weight/line_height/combinations), spacing, border_radius, shadows, component_styles (button/card/input), breakpoints.
 
@@ -117,7 +119,7 @@ Spawn agent to extract structured tokens from MASTER.md into `${PHASE_DIR}/desig
 
 #### 4c. Generate animation-tokens.json
 
-Spawn agent to generate `${PHASE_DIR}/design-ref/animation-tokens.json` from design-tokens + MASTER.md.
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Spawn agent to generate `${PHASE_DIR}/design-ref/animation-tokens.json` from design-tokens + MASTER.md.
 
 **Token schema keys:** duration (instant-slower), easing (ease-out/ease-in-out/spring), transitions, keyframes, interactions (hover states), reduced_motion (prefers-reduced-motion strategy).
 
@@ -140,7 +142,7 @@ Copy `design-system/MASTER.md` and `design-system/pages/*.md` (if generated) to
 
 #### 4f. Optional: Generate HTML prototype
 
-For each target, spawn agent to create `${PHASE_DIR}/design-ref/prototypes/${target}.html` from design-tokens + animation-tokens + MASTER.md.
+For each target, MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: spawn agent to create `${PHASE_DIR}/design-ref/prototypes/${target}.html` from design-tokens + animation-tokens + MASTER.md.
 
 **Rules:** standalone HTML with embedded CSS, realistic content (not lorem ipsum), SVG icons (Heroicons/Lucide CDN), cursor-pointer on clickables, responsive (375/768/1024px), WCAG AA, prefers-reduced-motion.
 
@@ -185,7 +187,7 @@ Same as ui-design.md — plan.md Step 4b detects `design-ref/MASTER.md` and incl
 | Error | Action |
 |-------|--------|
 | ui-ux-pro-max returns empty | Retry with broader keywords, then abort |
-| Token extraction agent fails | Retry once, warn if still fails |
+| Token extraction agent fails | Retry once, warn if still fails; if still fails, mark design-tokens.json as [LOW CONFIDENCE] (token extraction agent failed) |
 | User cancels selection | Save all variants, exit without solidification |
 
 ---

package/workflows/verify.md

@@ -405,7 +405,7 @@ For each success criterion / must-have truth:
 
 ### Step 4: Spawn Auditor Agent (if gaps found)
 
-Spawn workflow-nyquist-auditor agent with gap list, test infrastructure, and phase context.
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Spawn workflow-nyquist-auditor agent with gap list, test infrastructure, and phase context.
 Agent generates missing tests and returns:
 - GAPS FILLED -> record new tests
 - PARTIAL -> record resolved, escalate remainder
@@ -540,10 +540,10 @@ Next steps:
 |-------|--------|
 | Phase directory not found | Abort: "Phase {phase} not found." |
 | No execution results | Abort: "No completed tasks found. Run /workflow:execute first." |
-| No summaries found | Warn, proceed with task file analysis only |
+| No summaries found | Warn, proceed with task file analysis only; flag verification.json as [LOW CONFIDENCE] (partial, summaries missing) |
 | Test framework not detected | Skip coverage calculation, warn user |
-| Coverage command fails | Log error, proceed with requirement mapping only |
-| Verifier agent fails | Retry once, then write partial verification.json |
+| Coverage command fails | Log error, proceed with requirement mapping only; flag validation.json coverage as [LOW CONFIDENCE] (partial, coverage command failed) |
+| Verifier agent fails | Retry once, then write partial verification.json; flag verification.json as [LOW CONFIDENCE] (partial, verifier agent failed) |
 
 ---
 

package/workflows/wiki-connect.md

@@ -39,6 +39,8 @@ Knowledge graph link discovery and health improvement. Analyzes the unified wiki
 
 Gather baseline in parallel via `maestro wiki list --json`, `health`, `orphans`, `hubs --top 10`.
 
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: graph operations (`orphans`, `hubs --top 10`) — use the `maestro wiki` CLI to compute in-degree and orphan sets.
+
 Working state: entry count by type, baseline health score, orphan IDs, hub in-degree counts.
 
 Apply `--scope` filter if provided — restrict all subsequent analysis to matching type.

package/workflows/wiki-digest.md

@@ -74,8 +74,8 @@ Synthesize what these entries collectively teach. Focus on the knowledge pattern
 
 ### Key Entries
 Top 3-5 most important entries by:
-- Hub score (in-degree from `maestro wiki hubs`)
-- Backlink count (from `maestro wiki backlinks <id>`)
+- MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Hub score (in-degree from `maestro wiki hubs`)
+- MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: Backlink count (from `maestro wiki backlinks <id>`)
 - Recency (recently updated entries weigh more)
 
 ### Gap Detection
@@ -93,7 +93,7 @@ Per-theme health adapted from wiki health formula (entries, connectivity, comple
 
 Search via `maestro wiki search` or parse `.workflow/specs/learnings.md` for keyword matches against each theme. Flag **unlinked insights** — learning entries matching a theme but not referenced by any wiki entry in that theme.
 
-If `learnings.md` not found, skip with W002 warning.
+If `learnings.md` not found, skip with W002 warning; mark digest cross-reference as [LOW CONFIDENCE] (learnings.md missing).
 
 ---
 
@@ -152,6 +152,8 @@ Produce `.workflow/knowhow/KNW-digest-{slug}-{YYYY-MM-DD}.md`:
 
 ---
 
+GATE Stage 6→7: gaps identified BEFORE issue routing; BLOCKED if no gaps identified in Stage 5
+
 ## Stage 7: Gap → Issue Routing (if --create-issues)
 
 For each knowledge gap from Stage 5: dedup against `.workflow/issues/issues.jsonl` (same theme + type). If new, append with `type: "knowledge-gap"`, `status: "open"`, `severity: "low"`, `source: "wiki-digest"`, `tags: ["knowledge-gap", "{theme-slug}"]`. Report created count.

package/workflows/wiki-manage.md

@@ -40,6 +40,8 @@ Complements `wiki-connect.md` (link discovery) and `wiki-digest.md` (synthesis)
 
 Run in parallel: `maestro wiki health`, `list --json`, `orphans`, `hubs --top 5`.
 
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: graph operations (`orphans`, `hubs --top 5`) — use the `maestro wiki` CLI to compute orphan sets and hub in-degree.
+
 ### Step 2: Render Dashboard
 
 Display: health score, entry counts by type, broken links, orphan count, top hubs. Include health status message and quick-action commands (`/wiki-connect --fix`, `/wiki-digest`, `/manage-wiki cleanup --fix`, `maestro wiki graph`).
@@ -72,6 +74,8 @@ If not `--json`: offer to view an entry by number selection.
 
 Gather baseline via `maestro wiki health`, `orphans --json`, `graph`.
 
+MANDATORY, NOT SUBSTITUTABLE by manual Read/Grep: graph operations (`orphans --json`, `graph`) — use the `maestro wiki` CLI to compute orphans and graph topology.
+
 ### Step 2: Categorize Issues
 
 | Issue Type | Detection | Auto-fix Action |

package/workflows/maestro.codex.md

@@ -1,365 +0,0 @@
-# Workflow: Maestro (Codex Edition)
-
----
-
-## Step 1: Parse Arguments
-
-Extract from `$ARGUMENTS`:
-- Flags: `-y`/`--yes` (AUTO_YES), `-c`/`--continue` (RESUME), `--dry-run`, `--chain <name>`
-- `intent` = remaining text after flag removal
-
-**Resume mode**: If RESUME, load latest `.workflow/.maestro/maestro-*/status.json`, set `current_step` to first pending step, jump to **Step 6**.
-
----
-
-## Step 2: Read Project State
-
-Read `.workflow/state.json` if present. Derive `projectState`:
-- `current_phase`: first in-progress execute artifact, else first phase without completed execute
-- Fields: `phase_slug`, `phase_status` (pending|exploring|planning|executing|verifying|testing|completed|blocked), `phase_artifacts`, `execution` (tasks_completed/total), `verification_status`, `review_verdict`, `uat_status`, `phases_total/completed`, `has_blockers`, `accumulated_context`
-
-If not initialized and no intent → Error E001.
-
----
-
-## Step 3: Classify Intent & Select Chain
-
-### 3a: Exact-match keywords (fast path)
-
-If `forceChain` is set → validate against chainMap and jump to **3c**.
-
-Exact-match keywords: `continue`/`next`/`go`/`继续`/`下一步` → `state_continue`; `status`/`状态`/`dashboard` → `status`. If matched, skip to **3c**.
-
-### 3a-2: Structured intent extraction (LLM-native)
-
-Instead of regex, extract a structured intent tuple using LLM semantic understanding:
-
-```json
-{
-  "action":    "<create|fix|analyze|plan|execute|review|test|debug|refactor|explore|manage|transition|continue|sync|learn|retrospect>",
-  "object":    "<feature|bug|issue|code|test|spec|phase|milestone|doc|performance|security|ui|memory|codebase|config>",
-  "scope":     "<module/file/area or null>",
-  "issue_id":  "<ISS-XXXXXXXX-NNN if mentioned, else null>",
-  "phase_ref": "<integer if mentioned, else null>",
-  "urgency":   "<low|normal|high>"
-}
-```
-
-**Key disambiguation**: "问题"/"issue"/"problem" as something broken → `object: "bug"` (routes to debug). As a tracked item (with ISS-ID or management context) → `object: "issue"` (routes to issue management). When ambiguous, prefer `"bug"`.
-
-### 3a-3: Route via action × object matrix
-
-Route via `action × object` matrix. If `issue_id` present → issue pipeline directly.
-
-| action | object-specific overrides | default |
-|--------|--------------------------|---------|
-| fix | bug/code/perf/security→debug, issue→issue | debug |
-| create | feature→quick, issue→issue, test→test_gen, spec→spec_generate, ui→ui_design, config→init | quick |
-| analyze | bug/code→analyze, issue→issue_analyze, codebase→spec_map | analyze |
-| explore | issue→issue_discover, feature/ui→brainstorm/ui_design | brainstorm |
-| plan | issue→issue_plan, spec→spec_generate | plan |
-| execute | issue→issue_execute | execute |
-| manage | issue→issue, milestone→milestone_audit, phase→phase_transition, memory/doc/codebase→memory/sync/codebase_refresh | status |
-| transition | phase→phase_transition, milestone→milestone_complete | phase_transition |
-| review, test, debug, refactor, continue, sync, learn, retrospect, release, amend, compose | — | self-named |
-
-**Clarity scoring**: 3 = action+object+scope, 2 = action+object, 1 = action only, 0 = empty.
-If `clarity < 2` and not `AUTO_YES`: request user input (max 2 rounds).
-
-### 3b: State-based routing (when `taskType === 'state_continue'`)
-
-Returns `{ chain, argsOverride? }`. Steps resolved from `chainMap[chain]`.
-
-| Condition | Chain |
-|-----------|-------|
-| Not initialized | `init` |
-| No phases, no roadmap, has accumulated_context | `next-milestone` (with deferred/decisions context) |
-| No phases | `brainstorm-driven` |
-| pending + has context | `plan` |
-| pending, no context | `analyze` |
-| exploring/planning + has plan | `execute` |
-| exploring/planning, no plan | `plan` |
-| executing, all tasks done | `review` |
-| executing, tasks remain | `execute` |
-| exec completed + no review | `review` |
-| exec completed + review BLOCK | `review-fix` |
-| exec completed + UAT pending | `test` |
-| exec completed + UAT passed | `milestone-close` |
-| exec completed + UAT failed | `debug` |
-| testing, UAT passed | `milestone-close` |
-| testing, UAT not passed | `debug` |
-| completed | `milestone-close` |
-| blocked | `debug` |
-| fallback | `status` |
-
-### 3c: Intent-based chain map
-
-```javascript
-const chainMap = {
-  // ── Single-step ──────────────────────────────────────────────────────────
-  'status':             [{ cmd: 'manage-status' }],
-  'init':               [{ cmd: 'maestro-init' }],
-  'analyze':            [{ cmd: 'maestro-analyze',        args: '{phase}' }],
-  'ui_design':          [{ cmd: 'maestro-impeccable',         args: '"{phase}" --chain build' }],
-  'plan':               [{ cmd: 'maestro-plan',            args: '{phase}' }],
-  'execute':            [{ cmd: 'maestro-execute',         args: '{phase}' }],
-  'test_gen':           [{ cmd: 'quality-auto-test',        args: '{phase}' }],
-  'auto_test':          [{ cmd: 'quality-auto-test',        args: '{phase}' }],
-  'test':               [{ cmd: 'quality-test',            args: '{phase}' }],
-  'debug':              [{ cmd: 'quality-debug',           args: '"{description}"' }],
-  'integration_test':   [{ cmd: 'quality-auto-test',       args: '{phase}' }],
-  'refactor':           [{ cmd: 'quality-refactor',        args: '"{description}"' }],
-  'review':             [{ cmd: 'quality-review',          args: '{phase}' }],
-  'retrospective':      [{ cmd: 'quality-retrospective',   args: '{phase}' }],
-  'learn':              [{ cmd: 'manage-learn',            args: '"{description}"' }],
-  'sync':               [{ cmd: 'quality-sync' }],
-  'phase_transition':   [{ cmd: 'maestro-milestone-audit' }, { cmd: 'maestro-milestone-complete' }],
-  'milestone_audit':    [{ cmd: 'maestro-milestone-audit' }],
-  'milestone_complete': [{ cmd: 'maestro-milestone-complete' }],
-  'codebase_rebuild':   [{ cmd: 'manage-codebase-rebuild' }],
-  'codebase_refresh':   [{ cmd: 'manage-codebase-refresh' }],
-  'spec_setup':         [{ cmd: 'spec-setup' }],
-  'spec_add':           [{ cmd: 'spec-add',                args: '"{description}"' }],
-  'spec_load':          [{ cmd: 'spec-load' }],
-  'spec_map':           [{ cmd: 'manage-codebase-rebuild' }],
-  'knowhow_capture':     [{ cmd: 'manage-knowhow-capture',   args: '"{description}"' }],
-  'knowhow':             [{ cmd: 'manage-knowhow',           args: '"{description}"' }],
-  'issue':              [{ cmd: 'manage-issue',            args: '"{description}"' }],
-  'issue_discover':     [{ cmd: 'manage-issue-discover',   args: '"{description}"' }],
-  'issue_analyze':      [{ cmd: 'maestro-analyze',          args: '--gaps "{description}"' }],
-  'issue_plan':         [{ cmd: 'maestro-plan',            args: '--gaps' }],
-  'issue_execute':      [{ cmd: 'maestro-execute',         args: '' }],
-  'quick':              [{ cmd: 'maestro-quick',           args: '"{description}"' }],
-  // ── Multi-step chains ────────────────────────────────────────────────────
-  'spec-driven': [
-    { cmd: 'maestro-init' },
-    { cmd: 'maestro-roadmap', args: '--mode full "{description}"' },
-    { cmd: 'maestro-plan',          args: '{phase}' },
-    { cmd: 'maestro-execute',       args: '{phase}' }
-  ],
-  'brainstorm-driven': [
-    { cmd: 'maestro-brainstorm', args: '"{description}"' },
-    { cmd: 'maestro-plan',       args: '{phase}' },
-    { cmd: 'maestro-execute',    args: '{phase}' }
-  ],
-  'impeccable-build': [
-    { cmd: 'maestro-impeccable', args: '"{phase}" --chain build' },
-    { cmd: 'maestro-plan',      args: '{phase}' },
-    { cmd: 'maestro-execute',   args: '{phase}' }
-  ],
-  'full-lifecycle': [
-    { cmd: 'maestro-plan',          args: '{phase}' },
-    { cmd: 'maestro-execute',       args: '{phase}' },
-    { cmd: 'quality-review',        args: '{phase}' },
-    { cmd: 'quality-test',          args: '{phase}' },
-    { cmd: 'maestro-milestone-audit' },
-    { cmd: 'maestro-milestone-complete' }
-  ],
-  'quality-loop': [
-    { cmd: 'quality-review',   args: '{phase}' },
-    { cmd: 'quality-test',     args: '{phase}' },
-    { cmd: 'quality-debug',    args: '--from-uat {phase}' },
-    { cmd: 'maestro-plan',     args: '{phase} --gaps' },
-    { cmd: 'maestro-execute',  args: '{phase}' }
-  ],
-  'milestone-close': [
-    { cmd: 'maestro-milestone-audit' },
-    { cmd: 'maestro-milestone-complete' }
-  ],
-  'roadmap-driven': [
-    { cmd: 'maestro-init' },
-    { cmd: 'maestro-roadmap',  args: '"{description}"' },
-    { cmd: 'maestro-plan',     args: '{phase}' },
-    { cmd: 'maestro-execute',  args: '{phase}' }
-  ],
-  'next-milestone': [
-    { cmd: 'maestro-roadmap',  args: '"{description}"' },
-    { cmd: 'maestro-plan',     args: '{phase}' },
-    { cmd: 'maestro-execute',  args: '{phase}' }
-  ],
-  'analyze-plan-execute': [
-    { cmd: 'maestro-analyze', args: '"{description}" -q' },
-    { cmd: 'maestro-plan',    args: '--dir {scratch_dir}' },
-    { cmd: 'maestro-execute', args: '--dir {scratch_dir}' }
-  ],
-
-  // ── SKILL.md simplified aliases (--chain <name> shortcuts) ───────────────
-  'feature': [
-    { cmd: 'maestro-plan',    args: '{phase}' },
-    { cmd: 'maestro-execute', args: '{phase}' }
-  ],
-  'quality-fix': [
-    { cmd: 'maestro-analyze',      args: '--gaps "{description}"' },
-    { cmd: 'maestro-plan',         args: '--gaps' },
-    { cmd: 'maestro-execute',      args: '' }
-  ],
-  'deploy': [
-    { cmd: 'maestro-execute',         args: '{phase}' },
-    { cmd: 'maestro-milestone-release' }
-  ],
-
-  // ── Issue lifecycle chains (with quality gates) ────────────────────────────
-  'issue-full': [
-    { cmd: 'maestro-analyze',      args: '--gaps {issue_id}' },
-    { cmd: 'maestro-plan',         args: '--gaps' },
-    { cmd: 'maestro-execute',      args: '' },
-    { cmd: 'quality-review',       args: '{phase}' },
-    { cmd: 'manage-issue',         args: 'close {issue_id} --resolution fixed' }
-  ],
-  'issue-quick': [
-    { cmd: 'maestro-plan',         args: '--gaps' },
-    { cmd: 'maestro-execute',      args: '' },
-    { cmd: 'manage-issue',         args: 'close {issue_id} --resolution fixed' }
-  ],
-
-  'review-fix': [
-    { cmd: 'maestro-plan',    args: '{phase} --gaps' },
-    { cmd: 'maestro-execute', args: '{phase}' },
-    { cmd: 'quality-review',  args: '{phase}' }
-  ],
-  'quality-loop-partial': [
-    { cmd: 'maestro-plan',    args: '{phase} --gaps' },
-    { cmd: 'maestro-execute', args: '{phase}' }
-  ],
-  'milestone-release': [
-    { cmd: 'maestro-milestone-audit' },
-    { cmd: 'maestro-milestone-release' }
-  ],
-
-  'learn':              [{ cmd: 'maestro-learn',            args: '"{description}"' }],
-  'harvest':            [{ cmd: 'manage-harvest',           args: '"{description}"' }],
-  'wiki':               [{ cmd: 'manage-wiki' }],
-  'wiki_connect':       [{ cmd: 'wiki-connect' }],
-  'wiki_digest':        [{ cmd: 'wiki-digest' }],
-  'business_test':      [{ cmd: 'quality-auto-test',        args: '{phase}' }],
-  'spec_remove':        [{ cmd: 'spec-remove',              args: '"{description}"' }],
-  'amend':              [{ cmd: 'maestro-amend',             args: '"{description}"' }],
-  'release':            [{ cmd: 'maestro-milestone-release' }],
-  'compose':            [{ cmd: 'maestro-composer',          args: '"{description}"' }],
-  'play':               [{ cmd: 'maestro-player',            args: '"{description}"' }],
-  'update':             [{ cmd: 'maestro-update' }],
-  'overlay':            [{ cmd: 'maestro-overlay',           args: '"{description}"' }],
-  'link_coordinate':    [{ cmd: 'maestro-link-coordinate',   args: '"{description}"' }],
-};
-
-// Aliases: task type → named chain
-const taskToChain = {
-  'spec_generate':  'spec-driven',
-  'brainstorm':     'brainstorm-driven',
-  'issue_execute':  'issue-full',    // issue execute always gets review gate
-};
-```
-
-**Resolution order:**
-1. `forceChain` → `chainMap[forceChain]` (E002 if not found)
-2. `state_continue` → `detectNextAction(projectState)` → returns `{ chain, argsOverride? }`. Steps from `chainMap[chain]`. If `argsOverride` present, apply before template substitution.
-3. `taskToChain[taskType]` → named chain
-4. `chainMap[taskType]` → direct lookup
-
-### 3d: Resolve phase, description, and issue ID
-
-**Phase**: from structured extraction → fallback regex (`phase N` or bare number) → `projectState.current_phase`.
-**Issue ID**: from structured extraction → regex match `ISS-*-NNN`.
-
-Build context: `{ current_phase, user_intent, issue_id, spec_session_id: null, scratch_dir: null }`.
-
----
-
-## Step 4: Confirm
-
-**If `DRY_RUN`**: Display chain and exit.
-
-```
-MAESTRO-COORDINATE: {chain_name}  (dry run)
-  1. ${step.skill} {step.args}
-  2. ${step.skill} {step.args}
-  …
-```
-
-**If not `AUTO_YES`**: Ask user via `functions.request_user_input`:
-- Execute all steps
-- Execute from step N
-- Cancel
-
----
-
-## Step 5: Setup Session
-
-Create session directory `.workflow/.maestro/maestro-{timestamp}/`.
-
-**Barrier skills** (solo wave, coordinator analyzes artifacts after): `maestro-analyze`, `maestro-plan`, `maestro-brainstorm`, `maestro-roadmap`, `maestro-execute`.
-
-**Auto-flag injection** (when AUTO_YES): `maestro-analyze/-brainstorm/-roadmap/-ui-design` → `-y`, `maestro-plan` → `--auto`, `quality-test` → `--auto-fix`, `quality-retrospective` → `--auto-yes`.
-
-Initialize `state.json` with: session_id, intent, chain_name, auto_yes, context (phase, dirs, issue_id, gaps), waves[], and steps[] (each with index, skill, args, status=pending).
-
----
-
-## Step 6: Wave Execution Loop
-
-### 6a: Helper functions
-
-**buildSkillCall**: Replace template placeholders (`{phase}`, `{description}`, `{issue_id}`, `{plan_dir}`, `{analysis_dir}`, `{brainstorm_dir}`, `{spec_session_id}`, `{scratch_dir}`) from context. Inject auto-flag if AUTO_YES. Return `$<cmd> <args>`.
-
-**buildNextWave**: Take first pending step. If barrier → solo wave. Otherwise group consecutive non-barrier steps into one wave.
-
-### 6b: Wave instruction template
-
-Sub-agent instruction: execute `{skill_call}`, complete `{topic}`, do not modify state files, call `report_agent_job_result` with result JSON.
-
-Result schema: `{ status: "completed"|"failed", skill_call, summary, artifacts, error }` (all required).
-
-### 6c: Main loop
-
-While pending steps remain:
-1. Build next wave via `buildNextWave` (barrier → solo, non-barriers → grouped)
-2. Write wave CSV (`id, skill_call, topic`) to session dir
-3. Execute via `spawn_agents_on_csv` (max_workers = wave size, timeout 1800s)
-4. Read results CSV, update each step's status/findings/artifacts
-5. If barrier wave → run `analyzeBarrierArtifacts` to update context
-6. Record wave in state, persist `state.json`
-7. On any failure → abort (mark remaining steps skipped, break)
-
----
-
-## Step 7: Barrier Artifact Analysis
-
-After a barrier skill completes, the coordinator reads its artifacts and updates `context` for subsequent waves:
-
-Context updates per barrier skill:
-
-| Barrier | Extracts |
-|---------|----------|
-| `maestro-analyze` | `analysis_dir`, gaps (from context.md markers), phase (if detected) |
-| `maestro-plan` | `plan_dir`, task/wave count (from plan.json) |
-| `maestro-brainstorm` | `brainstorm_dir` |
-| `maestro-roadmap` | `spec_session_id` (SPEC-* pattern) |
-| `maestro-execute` | `exec_completed`/`exec_failed` counts (from results.csv) |
-
-**Key principle**: The coordinator owns all context assembly. Sub-agents receive a fully-resolved `skill_call`.
-
----
-
-## Step 8: Completion Report
-
-Finalize `state.json` (status: completed or current, completed_at timestamp).
-
-Generate `context.md` report: session ID, chain name, waves executed, steps completed/failed, per-wave result table with context updates.
-
-Display completion banner: session, chain, wave results (per-step status + summary), artifacts path, resume command.
-
----
-
-## Core Rules
-
-1. **Semantic routing**: LLM-native structured extraction (`action × object`) replaces regex; disambiguates "问题" by context
-2. **Wave-by-wave**: Never start wave N+1 before wave N results are read and barrier artifacts analyzed
-3. **Barrier = solo wave**: A barrier skill always executes alone; coordinator analyzes its artifacts before proceeding
-4. **Non-barriers can parallel**: Consecutive non-barrier skills share a wave with `max_workers = N`
-5. **Coordinator owns context**: Sub-agents receive fully-resolved `skill_call` — no context discovery needed
-6. **Simple instruction**: Sub-agent instruction is minimal — "execute {skill_call}, report result"
-7. **Quality gates**: Issue chains auto-include review; `issue-full` is default for issue execution
-8. **report_agent_job_result**: Every agent MUST call this with the output schema
-9. **State.json tracks waves**: Each wave recorded with step IDs and results; `--continue` resumes from next pending
-10. **Dry-run is read-only**: Display chain with [BARRIER] markers, no execution
-11. **Abort on failure**: Failed step → skip remaining → report