Scout

Codex compatibility note:

• Invoke repository skills with $skill-name in Codex; this mirrored copy rewrites legacy Claude /skill-name references.
• Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
• User-question prompts mean to ask the user directly in Codex.
• Ignore Claude-specific mode-switch instructions when they appear.
• Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
• Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required spawn_agent subagent(s) for that task.
• Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
• For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
• If a required step/tool cannot run in this environment, stop and ask the user before adapting.

Codex Project-Reference Loading (No Hooks)

Codex uses static project-reference loading instead of runtime-injected project docs. When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.

Always read:

• docs/project-config.json (project-specific paths, commands, modules, and workflow/test settings)
• docs/project-reference/docs-index-reference.md (routes to the full docs/project-reference/* catalog)
• docs/project-reference/lessons.md (always-on guardrails and anti-patterns)

Missing/stale context route: If docs/project-config.json, the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any task-required reference doc is missing or stale, auto-run $project-init or the narrow setup route ($project-config, $docs-init, $scan-all, $scan --target=<key>, $claude-md-init) before ordinary project-specific work. If Codex mirrors or AGENTS.md are missing/stale, ask the user to run $sync-codex; do not auto-run it.

Situation-based docs:

• Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra): project-structure-reference.md
• Backend/CQRS/API/domain/entity changes: backend-patterns-reference.md, domain-entities-reference.md
• Frontend/UI/styling/design-system: frontend-patterns-reference.md, scss-styling-guide.md, design-system/README.md
• Spec authoring, docs/specs/ pathing, or TC format: feature-spec-reference.md, spec-system-reference.md, spec-principles.md
• Behavior/public-contract changes or spec-test-code sync: workflow-spec-test-code-cycle-reference.md plus the spec docs above
• Derived spec indexes/ERDs/reimplementation guides: spec-system-reference.md and source Feature Specs under docs/specs/
• Integration test implementation/review: integration-test-reference.md
• E2E test implementation/review: e2e-test-reference.md
• Code review/audit work: code-review-rules.md plus domain docs above based on changed files

Do not read all docs blindly. Start from docs-index-reference.md, then open only relevant files for the task.

[BLOCKING] Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval. [BLOCKING] Before each step or sub-skill call, update task tracking: set in_progress when step starts, set completed when step ends. [BLOCKING] Every completed/skipped step MUST include brief evidence or explicit skip reason. [BLOCKING] If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions.

Quick Summary

Goal: Deliver a complete, prioritized map of every file relevant to the task via fast, parallel codebase discovery — grep + graph combined — so downstream work starts with full coverage and zero blind spots.

Summary:

• Main step pipeline (run in order, never merge): Phase 0 classify scope → Step 1 analyze request (extract entities/keywords/file-types) → Step 2 spawn parallel scout subagents (grep/glob) → Step 3 graph-expand 2-3 key files (YOU, main agent) → Step 4 low-result check (<5 files → broader re-pass) → Step 5 synthesize numbered prioritized list → ask the user directly for next step.
• Classify scope FIRST (Phase 0: backend/frontend/full-stack) so you spawn only the agents you need — never the default 3 when the prompt names one layer.
• Sub-agents do the parallel grep/glob; only YOU (main agent) run the graph commands afterward — graph expansion on 2-3 key files is MANDATORY when .code-graph/graph.db exists and is the step that finds what grep can't.
• This is discovery, not analysis: return prioritized file paths fast (3-5 min), no content deep-dives — that's investigate's job.
• If <5 files surface, re-examine keywords and run a second pass with broader synonyms before synthesizing; close by asking the user (ask the user directly) which next step to take.

Workflow:

1. Phase 0: Classify — Detect search scope (backend/frontend/both) + keyword type
2. Analyze Request — Extract entity names, feature keywords, file types from prompt
3. Parallel Search — Spawn agents searching backend core, backend infra, frontend paths
4. Graph Expand (MANDATORY — DO NOT SKIP) — MUST ATTENTION run graph commands on 2-3 key files. Graph reveals complete dependency network grep CANNOT find.
5. Low-Result Check — If <5 files returned, re-examine keywords and run second pass
6. Synthesize — Combine grep + graph into numbered, prioritized file list

Key Rules:

• Speed over depth — return file paths only, no content analysis
• Target 3-5 minutes total; 3-minute timeout per agent
• NEVER skip graph expansion when .code-graph/graph.db exists
• --ext / --engine=external switches the search engine to the scout-external agent (gemini/opencode CLIs); default is internal subagents

Scout — Fast Codebase File Discovery

Phase 0: Classify Search Scope

Before spawning agents, classify request:

Think: Prompt mention specific layer? Entity exist in backend, frontend, or both? Adjust agent count — avoid spawning unnecessary agents.

When to Use

• Quickly locating relevant files across large codebase
• Beginning work on features spanning multiple directories
• Before changes affecting multiple parts
• Mapping file landscape before investigation or implementation

NOT for: Deep code analysis (use investigate), debugging (use debug-investigate), implementation (use workflow-feature).

Workflow

Step 1: Analyze Search Request

Extract from USER_PROMPT:

• Entity names (User, Customer, Order)
• Feature names (authentication, notification)
• File types needed (backend, frontend, or both)

Step 2: Execute Parallel Search

Spawn SCALE number of scout subagents in parallel via spawn_agent tool (agent_type: "scout").

WHY scout not Explore: Custom scout agents read .claude/agents/scout.md — includes graph CLI knowledge + Bash access. Built-in Explore agents have NO graph awareness.

Engine Selection (--ext / --engine=external)

Detect the engine flag in args (default: internal):

• Internal (default): spawn agent_type: "scout" — the parallel grep/glob/graph search described below.
• External (--ext or --engine=external): spawn agent_type: "scout-external" instead. That agent (.claude/agents/scout-external.md) owns the gemini/opencode CLI dispatch, the Explore fallback, and the install prompt when those CLIs are absent.

Flag only switches which subagent runs. Orchestration identical for both engines: Phase 0 classify, Step 3 graph-expand (run by you, main agent), low-result check, synthesize. Output contract (numbered, prioritized file list) same.

Agent Distribution

• Agent 1 - Backend Core: {module-source-root}/ domain folder + command folder + query folder (per the project's structure reference)
• Agent 2 - Backend Infra: {module-source-root}/ event-handler folder + controllers + background-jobs folder (per the project's structure reference)
• Agent 3 - Frontend: {frontend-apps-dir}/, {frontend-libs-dir}/{domain-lib}/, {frontend-libs-dir}/{common-lib}/

Per agent: 3-minute timeout. Return file paths only — no content analysis. Use Glob (patterns), Grep (content), Bash (graph CLI).

Search-Axis Fan-Out (ONE wave — each axis is blind to the others)

Partitioning by path still searches ONE way. Each axis below finds files the others structurally CANNOT see — that blindness is exactly why a single search angle misses files. Select the axes the prompt warrants (they compose with the path split above: an agent gets one path scope + one axis) and spawn the whole set TOGETHER in one message:

Graph is the sixth axis and it stays with YOU — Step 3, after the barrier; sub-agents do not run it here.

Wave rules: precompute the full path × triggered-axis assignment set, then dispatch it in consecutive capacity-bounded waves with one unique exact artifact path per assignment. Never drop an axis to fit host capacity; queue overflow assignments into the next wave. Each wave is read-only with zero write overlap, launches in one message, and reaches its barrier before Step 3. An axis the prompt gives no trigger for is not an assignment and counts as returned. Both engines take identical axis briefs — --ext only swaps scout for scout-external.

Step 3: Graph Expand (MANDATORY — DO NOT SKIP)

YOU (main agent) MUST ATTENTION run graph commands YOURSELF after sub-agents return. NOT optional — without graph, results are incomplete. Sub-agents cannot use graph — only main agent can.

# Check graph exists
ls .code-graph/graph.db 2>/dev/null && echo "GRAPH_AVAILABLE" || echo "NO_GRAPH"

If GRAPH_AVAILABLE, pick 2-3 key files from sub-agent results (entities, commands, bus messages):

# Full dependency network of key file
python .claude/scripts/code_graph connections <key_file> --json

# All callers of key command/handler
python .claude/scripts/code_graph query callers_of <FunctionName> --json

# All importers of bus message class
python .claude/scripts/code_graph query importers_of <file_path> --json

# Batch query multiple files (most efficient)
python .claude/scripts/code_graph batch-query <file1> <file2> <file3> --json

# If graph returns "ambiguous" — disambiguate first
python .claude/scripts/code_graph search <keyword> --kind Function --json

# Trace shortest path between two nodes
python .claude/scripts/code_graph find-path <source_qn> <target_qn> --json

# Filter by service, limit results
python .claude/scripts/code_graph query callers_of <name> --limit 5 --filter "ServiceName" --json

Grep-First Discovery (semantic queries): When prompt describes behavior/flow (not specific file), grep key terms FIRST to discover entry files, then use those as graph input:

1. Grep class names, commands, handlers, endpoints
2. Use discovered files as input to connections, batch-query, or trace
3. Use trace --direction both on middle files (controllers, commands) for full upstream + downstream

Graph results get HIGHER priority than grep — structural relationships > text matches. After graph expansion, grep again to verify content in discovered files.

Post-Grep Trace Trigger: whenever a grep/glob surfaces an important entry-point file — an entity, command, query, event/command handler, controller, bus message/consumer, component, store, or api-service — immediately run a graph trace on it before concluding. The trace reveals callers, consumers, bus messages, event chains, and tests that grep CANNOT find: python .claude/scripts/code_graph trace <key-entry-file> --direction both --json. Pattern: grep finds files → graph trace reveals full system flow → grep verifies specific details.

Step 4: Low-Result Check

If total files found <5 after Steps 2-3:

1. Re-examine keywords — too specific? Try broader synonyms
2. Spawn second scout agent with alternate search terms
3. Try python .claude/scripts/code_graph search <keyword> --json to find nodes by name

Step 5: Synthesize Results

De-duplicate the union FIRST. The axes overlap by design, so the same file legitimately arrives from several agents. Normalize each path (separators, casing) and merge to exactly ONE entry, recording which axes surfaced it — multi-axis hits rank higher (independent confirmation), while a HIGH-priority file surfaced by a single axis is the coverage risk worth one verifying grep before it ships in the list.

Then combine grep + graph into a numbered, prioritized file list (see Results Format).

Search Patterns by Priority

Substitute folder names + file globs from the project's structure reference / docs/project-config.json. {backend-source-glob} / {frontend-source-glob} are the per-stack source extensions.

# HIGH PRIORITY - Core Logic
**/{entity-folder}/**/*{keyword}*.{backend-source-glob}
**/{command-folder}/**/*{keyword}*.{backend-source-glob}
**/{query-folder}/**/*{keyword}*.{backend-source-glob}
**/{event-handler-folder}/**/*{keyword}*.{backend-source-glob}
**/*{keyword}*{component-suffix}.{frontend-source-glob}
**/*{keyword}*{store-suffix}.{frontend-source-glob}

# MEDIUM PRIORITY - Infrastructure
**/{controllers-folder}/**/*{keyword}*.{backend-source-glob}
**/{background-jobs-folder}/**/*{keyword}*.{backend-source-glob}
**/*Consumer*{keyword}*.{backend-source-glob}
**/*{keyword}*{api-service-suffix}.{frontend-source-glob}

# LOW PRIORITY - Supporting
**/*{keyword}*Helper*.{backend-source-glob}
**/*{keyword}*Service*.{backend-source-glob}
**/*{keyword}*{markup-glob}

Graph Intelligence (MANDATORY when graph.db exists)

Results Format

## Scout Results: {USER_PROMPT}

### High Priority - Core Logic

1. `{module-source-root}/{entity-folder}/{Entity}` — domain entity
2. `{module-source-root}/{command-folder}/{Feature}/Save{Entity}Command` — mutating command
   ...

### Medium Priority - Infrastructure

10. `{module-source-root}/{controllers-folder}/{Entity}Controller` — endpoint
11. `{module-source-root}/{event-handler-folder}/{Feature}/SendNotificationOn{Entity}CreatedEventHandler` — event handler
    ...

### Low Priority - Supporting

20. `{module-source-root}/{helpers-folder}/{Entity}Helper` — supporting helper
    ...

### Frontend Files

30. `{frontend-libs-dir}/{domain-lib}/{configured-feature-path}/{feature}-list.component`
    ...

**Total Files Found:** {count}
**Search Completed In:** {time}

### Suggested Starting Points

1. `{most relevant file}` - {reason}
2. `{second most relevant}` - {reason}

### End-to-Start Trace Candidates

| Role                           | Candidate files | Why relevant                                           | Evidence                         |
| ------------------------------ | --------------- | ------------------------------------------------------ | -------------------------------- |
| Observed final output / reader | `{files}`       | `{reader, renderer, assertion, query, aggregate, log}` | `{file:line or search evidence}` |
| Storage / projection / cache   | `{files}`       | `{state consumed by reader}`                           | `{file:line or search evidence}` |
| Writer / updater               | `{files}`       | `{writes final state}`                                 | `{file:line or search evidence}` |
| Consumer / handler / job       | `{files}`       | `{transforms or schedules writes}`                     | `{file:line or search evidence}` |
| Producer / origin trigger      | `{files}`       | `{upstream source of input}`                           | `{file:line or search evidence}` |

**Feeder-path scan:** list every producer/caller/event/job candidate that may write the same final state. Mark unknown paths explicitly instead of hiding them.

### Unresolved Questions

- {any questions that need clarification}

Quality Standards

Workflow Recommendation

MANDATORY MUST ATTENTION — NO EXCEPTIONS: If NOT already in workflow, MUST ATTENTION use ask the user directly to ask user:

1. Activate investigation workflow (Recommended) — scout → investigate
2. Execute $scout directly — run this skill standalone

Next Steps

MANDATORY MUST ATTENTION — NO EXCEPTIONS after completing, MUST ATTENTION use ask the user directly to present:

• "$investigate (Recommended)" — Deep-dive into discovered files
• "$plan" — If scouted files sufficient to start planning
• "Skip, continue manually" — user decides

[IMPORTANT] Use task tracking to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ATTENTION ask user whether to skip.

• docs/project-reference/domain-entities-reference.md — Domain entity catalog, relationships, cross-service sync (read when task involves business entities/models)

External Memory: Complex/lengthy work → write findings incrementally to plans/reports/. Prevents context loss.

Evidence Gate: MANDATORY MUST ATTENTION — every claim, finding, recommendation requires file:line proof with confidence % (>80% act, <80% verify first).

Graph-Assisted Investigation — MANDATORY when .code-graph/graph.db exists.

HARD-GATE: MUST ATTENTION run at least ONE graph command on key files before concluding any investigation.

Pattern: Grep finds files → trace --direction both reveals full system flow → Grep verifies details

Task	Minimum Graph Action
Investigation/Scout	trace --direction both on 2-3 entry files
Fix/Debug	callers_of on buggy function + tests_for
Feature/Enhancement	connections on files to be modified
Code Review	tests_for on changed functions
Blast Radius	trace --direction downstream

CLI: python .claude/scripts/code_graph {command} --json. Use --node-mode file first (10-30x less noise), then --node-mode function for detail.

Incremental Result Persistence — MANDATORY for all sub-agents or heavy inline steps processing >3 files.

1. Before starting: Create report file plans/reports/{skill}-{date}-{slug}.md
2. After each file/section reviewed: Append findings to report immediately — never hold in memory
3. Return to main agent: Summary only (per SYNC:subagent-return-contract) with Full report: path
4. Main agent: Reads report file only when resolving specific blockers

Why: Context cutoff mid-execution loses ALL in-memory findings. Each disk write survives compaction. Partial results are better than no results.

Report naming: plans/reports/{skill-name}-{YYMMDD}-{HHmm}-{slug}.md

Sub-Agent Return Contract — When this skill spawns a sub-agent, the sub-agent MUST return ONLY this structure. Main agent reads only this summary — NEVER requests full sub-agent output inline.

## Sub-Agent Result: [skill-name]

Status: ✅ PASS | ⚠️ PARTIAL | ❌ FAIL
Confidence: [0-100]%

### Findings (Critical/High only — max 10 bullets)

- [severity] [file:line] [finding]

### Actions Taken

- [file changed] [what changed]

### Blockers (if any)

- [blocker description]

Full report: plans/reports/[skill-name]-[date]-[slug].md

Main agent reads Full report file ONLY when: (a) resolving a specific blocker, or (b) building a fix plan. Sub-agent writes full report incrementally (per SYNC:incremental-persistence) — not held in memory.

Context budget — the return payload is a SUMMARY, not a transcript: ≤10 finding bullets, no raw file contents / full diffs / verbatim logs inline, no re-pasted source. Everything beyond the summary lives in the Full report on disk. A sub-agent that would exceed the summary shape MUST write the detail to its report and return only the pointer — the orchestrator's context is the scarce resource the whole map-reduce protects.

Nested Task Expansion Contract — For workflow-step invocation, the [Workflow] ... row is only a parent container; the child skill still creates visible phase tasks.

1. Call the current task list first. If a matching active parent workflow row exists, set nested=true and record parentTaskId; otherwise run standalone.
2. Create one task per declared phase before phase work. When nested, prefix subjects [N.M] $skill-name — phase.
3. When nested, link the parent with TaskUpdate(parentTaskId, addBlockedBy: [childIds]).
4. Orchestrators must pre-expand a child skill's phase list and link the workflow row before invoking that child skill or sub-agent.
5. Mark exactly one child in_progress before work and completed immediately after evidence is written.
6. Complete the parent only after all child tasks are completed or explicitly cancelled with reason.

Blocked until: the current task list done, child phases created, parent linked when nested, first child marked in_progress.

Project Reference Docs Gate — Run after task-tracking bootstrap and before target/source file reads, grep, edits, or analysis. Project docs override generic framework assumptions.

1. Identify scope: file types, domain area, and operation.
2. Read docs/project-config.json first — the project's machine-readable map. It is the single source of truth for THIS repo (modules/paths, framework + search keywords, test/E2E/integration run-commands, design system, architecture rules, workflow patterns); ground exact paths, run-commands, and conventions on it before investigating, planning, or coding — never assume framework defaults (CLAUDE.md + reference docs are derived from it). If it — or the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any required reference doc — is missing or stale, auto-run $project-init or the narrow route ($project-config, $docs-init, $scan-all, $scan --target=<key>, $claude-md-init) first; if Codex mirrors or AGENTS.md are stale, ask the user to run $sync-codex (never auto-run it).
3. Required docs by trigger: always docs/project-reference/lessons.md; doc lookup docs-index-reference.md; review code-review-rules.md; backend/CQRS/API backend-patterns-reference.md; domain/entity domain-entities-reference.md; frontend/UI frontend-patterns-reference.md; styles/design scss-styling-guide.md + design-system/design-system-canonical.md; integration tests integration-test-reference.md; E2E e2e-test-reference.md; feature docs/specs feature-spec-reference.md + spec-system-reference.md + spec-principles.md; behavior/public-contract/spec-test-code sync workflow-spec-test-code-cycle-reference.md; derived spec index/ERD/reimplementation guides spec-system-reference.md + source Feature Specs under docs/specs/; architecture/new area project-structure-reference.md.
4. Read every required doc, then before target work state: Reference docs read: ... | Not applicable: ....

Ready when: scope evaluated, docs/project-config.json consulted, required docs checked/read or setup route completed, lessons.md confirmed, citation emitted.

Task Tracking & External Report Persistence — Bootstrap this before execution; then run project-reference doc prefetch before target/source work.

1. Create a small task breakdown before target file reads, grep, edits, or analysis. On context loss, inspect the current task list first.
2. Mark one task in_progress before work and completed immediately after evidence; never batch transitions.
3. For plan/review work, create plans/reports/{skill}-{YYMMDD}-{HHmm}-{slug}.md before first finding.
4. Append findings after each file/section/decision and synthesize from the report file at the end.
5. Final output cites Full report: plans/reports/{filename}.

Blocked until: task breakdown exists, report path declared for plan/review work, first finding persisted before the next finding.

Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.

Evidence-Based Reasoning — Speculation is FORBIDDEN. Every claim needs proof.

1. Cite file:line, grep results, or framework docs for EVERY claim
2. Declare confidence: >80% act freely, 60-80% verify first, <60% DO NOT recommend
3. Cross-service validation required for architectural changes
4. "I don't have enough evidence" is valid and expected output

BLOCKED until: - [ ] Evidence file path (file:line) - [ ] Grep search performed - [ ] 3+ similar patterns found - [ ] Confidence level stated

Forbidden without proof: "obviously", "I think", "should be", "probably", "this is because" If incomplete → output: "Insufficient evidence. Verified: [...]. Not verified: [...]."

Cross-Service Check — Microservices/event-driven: MANDATORY before concluding investigation, plan, spec, or feature doc. Missing downstream consumer = silent regression.

Boundary	Grep terms
Event producers	Publish, Dispatch, Send, emit, EventBus, outbox, IntegrationEvent
Event consumers	Consumer, EventHandler, Subscribe, @EventListener, inbox
Sagas/orchestration	Saga, ProcessManager, Choreography, Workflow, Orchestrator
Sync service calls	HTTP/gRPC calls to/from other services
Shared contracts	OpenAPI spec, proto, shared DTO — flag breaking changes
Data ownership	Other service reads/writes same table/collection → Shared-DB anti-pattern

Per touchpoint: owner service · message name · consumers · risk (NONE / ADDITIVE / BREAKING).

BLOCKED until: Producers scanned · Consumers scanned · Sagas checked · Contracts reviewed · Breaking-change risk flagged

Rationalization Prevention — AI skips steps via these evasions. Recognize and reject:

Evasion	Rebuttal
"Too simple for a plan"	Simple + wrong assumptions = wasted time. Plan anyway.
"I'll test after"	RED before GREEN. Write/verify test first.
"Already searched"	Show grep evidence with file:line. No proof = no search.
"Just do it"	Still need task tracking. Skip depth, never skip tracking.
"Just a small fix"	Small fix in wrong location cascades. Verify file:line first.
"Code is self-explanatory"	Future readers need evidence trail. Document anyway.
"Combine steps to save time"	Combined steps dilute focus. Each step has distinct purpose.

AI Mistake Prevention — Failure modes to avoid on every task:

Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting. Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing. Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first. Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done. Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect. Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard. Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk. Assert the outcome your system owns, not the intermediate state your infrastructure owns. When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure. Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.

MUST ATTENTION cite file:line evidence for every claim. Confidence >80% to act, <60% = do NOT recommend.

MUST ATTENTION never skip steps via evasions. Plan anyway. Test first. Show grep evidence with file:line.

MUST ATTENTION run at least ONE graph command on key files before concluding when .code-graph/graph.db exists.

MUST ATTENTION apply critical + sequential thinking — every claim needs appropriate traced evidence (file:line for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.

MUST ATTENTION apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.

• MANDATORY Bootstrap task tracking before target work; transition one task at a time.
• MANDATORY Persist plan/review findings to plans/reports/ incrementally and synthesize from disk.

• MANDATORY Before investigating, planning, or coding, read docs/project-config.json (the project map: modules/paths, run-commands, conventions, architecture/workflow rules) + the required project-reference docs, and cite Reference docs read: ....
• MANDATORY Always include lessons.md; project config + conventions override generic framework defaults.
• MANDATORY If project config, root instruction files, or any required reference doc is missing or stale, auto-run $project-init or the narrow lower-level route before ordinary project-specific work.

• MANDATORY Parent workflow rows do not replace child phase tracking; expand phases and link the parent when nested.
• MANDATORY Orchestrators pre-expand child skill phases before invocation; use [N.M] $skill-name — phase prefixes and one-in_progress discipline.

Prompt-Enhance Closing Anchors

IMPORTANT MUST ATTENTION follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval IMPORTANT MUST ATTENTION for every step/sub-skill call: set in_progress before execution, set completed after execution IMPORTANT MUST ATTENTION every skipped step MUST include explicit reason; every completed step MUST include concise evidence IMPORTANT MUST ATTENTION if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses

Parallel Sub-Agent Dispatch — Plan parallelism the moment a task breakdown exists, BEFORE executing it — running provably independent tasks sequentially wastes wall-clock. Applies to every multi-step job: workflow steps, planning, batch updates, investigation, research, scans, reviews, doc sync. Plan execution is metadata-gated, NEVER default-parallel — fan-out follows ONLY what the plan declares (PAR/SEQ tags + per-phase write set); an untagged plan runs sequentially — why: a derived write set cannot see cascade or generated writes.

1. Tag every task PAR or SEQ. PAR = inputs exclude every pending task's output AND write set disjoint from every other PAR. Else SEQ — MUST ATTENTION name the dependency forcing it.
2. Group PAR into waves. No edge between members. Two writers of one file NEVER share a wave. Read-only work (search, investigation, review, research) parallelizes freely.
3. Declare before dispatch: Parallel plan: wave 1 = [...] · wave 2 = [...] · SEQ = [...] (reason).
4. Spawn each wave in ONE message — every spawn_agent call in one response, NEVER dripped per turn. Route each task to its specialist (.claude/skills/shared/sub-agent-selection-guide.md); NEVER code-reviewer as catch-all.
5. Brief each sub-agent self-contained: goal · scope + owned files · reference docs · return contract (summary + Full report: path, per SYNC:subagent-return-contract) · incremental persistence to plans/reports/ (per SYNC:incremental-persistence).
6. Barrier per wave. Advance ONLY after EVERY member returns (a skipped conditional counts as returned). Merge, mark each task completed/skipped, THEN dispatch the next wave. Mutating steps wait for the barrier.
7. One level deep. A dispatched sub-agent executes its own brief; further fan-out stays the orchestrator's job unless that agent's .claude/agents/*.md definition authorizes it.

NEVER parallelize: tasks sharing a write target · a task consuming a pending task's output · trivial single-file work (dispatch overhead > gain) · an order a skill or workflow explicitly fixes · gates awaiting user approval.

Blocked until: MUST ATTENTION every task tagged PAR/SEQ with a named reason per SEQ · waves declared + write-set disjointness checked · each wave spawned in ONE message · barrier honored before the next wave.

• MANDATORY After planning tasks, tag each PAR/SEQ and spawn every PAR wave as parallel sub-agents in ONE message — default parallel for workflows, batch updates, investigation, research, reviews; plan execution fans out ONLY on what the plan declares.
• MANDATORY Disjoint write sets per wave · all-return barrier before the next wave · specialist routing · sub-agents NEVER fan out further unless their own agent definition authorizes it.

Project Protocol Overlay — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this sk

…(truncated)