cook（cook 数据库）

[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.

[IMPORTANT] Use TaskCreate 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.

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.

AI Mistake Prevention — Failure modes to avoid on every task: - Check downstream references before deleting. Deleting components causes documentation and code staleness cascades. Map all referencing files before removal. - Verify AI-generated content against actual code. AI hallucinates APIs, class names, and method signatures. Always grep to confirm existence before documenting or referencing. - Trace full dependency chain after edits. Changing a definition misses downstream variables and consumers derived from it. Always trace the full chain. - Trace ALL code paths when verifying correctness. Confirming code exists is not confirming it executes. Always trace early exits, error branches, and conditional skips — not just happy path. - When debugging, ask "whose responsibility?" before fixing. Trace whether bug is in caller (wrong data) or callee (wrong handling). Fix at responsible layer — never patch symptom site. - Assume existing values are intentional — ask WHY before changing. Before changing any constant, limit, flag, or pattern: read comments, check git blame, examine surrounding code. - Verify ALL affected outputs, not just the first. Changes touching multiple stacks require verifying EVERY output. One green check is not all green checks. - Holistic-first debugging — resist nearest-attention trap. When investigating any failure, list EVERY precondition first (config, env vars, DB names, endpoints, DI registrations, data preconditions), then verify each against evidence before forming any code-layer hypothesis. - Surgical changes — apply the diff test. Bug fix: every changed line must trace directly to the bug. Don't restyle or improve adjacent code. Enhancement task: implement improvements AND announce them explicitly. - Surface ambiguity before coding — don't pick silently. If request has multiple interpretations, present each with effort estimate and ask. Never assume all-records, file-based, or more complex path.

Understand Code First — HARD-GATE: Do NOT write, plan, or fix until you READ existing code. 1. Search 3+ similar patterns (grep/glob) — cite file:line evidence 2. Read existing files in target area — understand structure, base classes, conventions 3. Run python.claude/scripts/code_graph trace <file> --direction both --json when .code-graph/graph.db exists 4. Map dependencies via connections or callers_of — know what depends on your target 5. Write investigation to .ai/workspace/analysis/ for non-trivial tasks (3+ files) 6. Re-read analysis file before implementing — never work from memory alone 7. NEVER invent new patterns when existing ones work — match exactly or document deviation BLOCKED until: - [] Read target files - [] Grep 3+ patterns - [] Graph trace (if graph.db exists) - [] Assumptions verified with evidence

Deep Multi-Round Review — Escalating rounds. Round 1 in main session. Round 2+ and EVERY recursive re-review iteration MUST use a fresh sub-agent. Round 1: Main-session review. Read target files, build understanding, note issues. Output baseline findings. Round 2: MANDATORY fresh sub-agent review — see SYNC:fresh-context-review for the spawn mechanism and SYNC:review-protocol-injection for the canonical Agent prompt template. The sub-agent re-reads ALL files from scratch with ZERO Round 1 memory. It must catch: - Cross-cutting concerns missed in Round 1 - Interaction bugs between changed files - Convention drift (new code vs existing patterns) - Missing pieces that should exist but don't - Subtle edge cases the main session rationalized away Round 3+ (recursive after fixes): After ANY fix cycle, MANDATORY fresh sub-agent re-review. Spawn a NEW Agent tool call each iteration — never reuse Round 2's agent. Each new agent re-reads ALL files from scratch with full protocol injection. Continue until PASS or 3 fresh-subagent rounds max, then escalate to user via AskUserQuestion. Rules: - NEVER declare PASS after Round 1 alone - NEVER reuse a sub-agent across rounds — every iteration spawns a NEW Agent call - Main agent READS sub-agent reports but MUST NOT filter, reinterpret, or override findings - Max 3 fresh-subagent rounds per review — if still FAIL, escalate via AskUserQuestion (do NOT silently loop) - Track round count in conversation context (session-scoped) - Final verdict must incorporate ALL rounds Report must include ## Round N Findings (Fresh Sub-Agent) for every round N≥2.

Fresh Sub-Agent Review — Eliminate orchestrator confirmation bias via isolated sub-agents. Why: The main agent knows what it (or /cook) just fixed and rationalizes findings accordingly. A fresh sub-agent has ZERO memory, re-reads from scratch, and catches what the main agent dismissed. Sub-agent bias is mitigated by (1) fresh context, (2) verbatim protocol injection, (3) main agent not filtering the report. When: Round 2 of ANY review AND every recursive re-review iteration after fixes. NOT needed when Round 1 already PASSes with zero issues. How: 1. Spawn a NEW Agent tool call — use code-reviewer subagent_type for code reviews, general-purpose for plan/doc/artifact reviews 2. Inject ALL required review protocols VERBATIM into the prompt — see SYNC:review-protocol-injection for the full list and template. Never reference protocols by file path; AI compliance drops behind file-read indirection (see SYNC:shared-protocol-duplication-policy) 3. Sub-agent re-reads ALL target files from scratch via its own tool calls — never pass file contents inline in the prompt 4. Sub-agent writes structured report to plans/reports/{review-type}-round{N}-{date}.md 5. Main agent reads the report, integrates findings into its own report, DOES NOT override or filter Rules: - NEVER reuse a sub-agent across rounds — every iteration spawns a NEW Agent call - NEVER skip fresh-subagent review because "last round was clean" — every fix triggers a fresh round - Max 3 fresh-subagent rounds per review — escalate via AskUserQuestion if still failing; do NOT silently loop or fall back to any prior protocol - Track iteration count in conversation context (session-scoped, no persistent files)

Review Protocol Injection — Every fresh sub-agent review prompt MUST embed 10 protocol blocks VERBATIM. The template below has ALL 10 bodies already expanded inline. Copy the template wholesale into the Agent call's prompt field at runtime, replacing only the {placeholders} in Task / Round / Reference Docs / Target Files / Output sections with context-specific values. Do NOT touch the embedded protocol sections. Why inline expansion: Placeholder markers would force file-read indirection at runtime. AI compliance drops significantly behind indirection (see SYNC:shared-protocol-duplication-policy). Therefore the template carries all 10 protocol bodies pre-embedded.

Subagent Type Selection

• code-reviewer — for code reviews (reviewing source files, git diffs, implementation)
• general-purpose — for plan / doc / artifact reviews (reviewing markdown plans, docs, specs)

Canonical Agent Call Template (Copy Verbatim)

Agent({
  description: "Fresh Round {N} review",
  subagent_type: "code-reviewer",
  prompt: `
## Task
{review-specific task — e.g., "Review all uncommitted changes for code quality" | "Review plan files under {plan-dir}" | "Review integration tests in {path}"}

## Round
Round {N}. You have ZERO memory of prior rounds. Re-read all target files from scratch via your own tool calls. Do NOT trust anything from the main agent beyond this prompt.

## Protocols (follow VERBATIM — these are non-negotiable)

### 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) provided; 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: [...]."

### Bug Detection
MUST check categories 1-4 for EVERY review. Never skip.
1. Null Safety: Can params/returns be null? Are they guarded? Optional chaining gaps? .find() returns checked?
2. Boundary Conditions: Off-by-one (< vs <=)? Empty collections handled? Zero/negative values? Max limits?
3. Error Handling: Try-catch scope correct? Silent swallowed exceptions? Error types specific? Cleanup in finally?
4. Resource Management: Connections/streams closed? Subscriptions unsubscribed on destroy? Timers cleared? Memory bounded?
5. Concurrency (if async): Missing await? Race conditions on shared state? Stale closures? Retry storms?
6. Stack-Specific: JS: === vs ==, typeof null. C#: async void, missing using, LINQ deferred execution.
Classify: CRITICAL (crash/corrupt) → FAIL | HIGH (incorrect behavior) → FAIL | MEDIUM (edge case) → WARN | LOW (defensive) → INFO.

### Design Patterns Quality
Priority checks for every code change:
1. DRY via OOP: Same-suffix classes (*Entity, *Dto, *Service) MUST share base class. 3+ similar patterns → extract to shared abstraction.
2. Right Responsibility: Logic in LOWEST layer (Entity > Domain Service > Application Service > Controller). Never business logic in controllers.
3. SOLID: Single responsibility (one reason to change). Open-closed (extend, don't modify). Liskov (subtypes substitutable). Interface segregation (small interfaces). Dependency inversion (depend on abstractions).
4. After extraction/move/rename: Grep ENTIRE scope for dangling references. Zero tolerance.
5. YAGNI gate: NEVER recommend patterns unless 3+ occurrences exist. Don't extract for hypothetical future use.
Anti-patterns to flag: God Object, Copy-Paste inheritance, Circular Dependency, Leaky Abstraction.

### Logic & Intention Review
Verify WHAT code does matches WHY it was changed.
1. Change Intention Check: Every changed file MUST serve the stated purpose. Flag unrelated changes as scope creep.
2. Happy Path Trace: Walk through one complete success scenario through changed code.
3. Error Path Trace: Walk through one failure/edge case scenario through changed code.
4. Acceptance Mapping: If plan context available, map every acceptance criterion to a code change.
NEVER mark review PASS without completing both traces (happy + error path).

### Test Spec Verification
Map changed code to test specifications.
1. From changed files → find TC-{FEAT}-{NNN} in docs/business-features/{Service}/detailed-features/{Feature}.md Section 15.
2. Every changed code path MUST map to a corresponding TC (or flag as "needs TC").
3. New functions/endpoints/handlers → flag for test spec creation.
4. Verify TC evidence fields point to actual code (file:line, not stale references).
5. Auth changes → TC-{FEAT}-02x exist? Data changes → TC-{FEAT}-01x exist?
6. If no specs exist → log gap and recommend /tdd-spec.
NEVER skip test mapping. Untested code paths are the #1 source of production bugs.

### Fix-Layer Accountability
NEVER fix at the crash site. Trace the full flow, fix at the owning layer. The crash site is a SYMPTOM, not the cause.
MANDATORY before ANY fix:
1. Trace full data flow — Map the complete path from data origin to crash site across ALL layers (storage → backend → API → frontend → UI). Identify where bad state ENTERS, not where it CRASHES.
2. Identify the invariant owner — Which layer's contract guarantees this value is valid? Fix at the LOWEST layer that owns the invariant, not the highest layer that consumes it.
3. One fix, maximum protection — If fix requires touching 3+ files with defensive checks, you are at the wrong layer — go lower.
4. Verify no bypass paths — Confirm all data flows through the fix point. Check for direct construction skipping factories, clone/spread without re-validation, raw data not wrapped in domain models, mutations outside the model layer.
BLOCKED until: Full data flow traced (origin → crash); Invariant owner identified with file:line evidence; All access sites audited (grep count); Fix layer justified (lowest layer that protects most consumers).
Anti-patterns (REJECT): "Fix it where it crashes" (crash site ≠ cause site, trace upstream); "Add defensive checks at every consumer" (scattered defense = wrong layer); "Both fix is safer" (pick ONE authoritative layer).

### Rationalization Prevention
AI skips steps via these evasions. Recognize and reject:
- "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 TaskCreate. 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.

### Graph-Assisted Investigation
MANDATORY when .code-graph/graph.db exists.
HARD-GATE: MUST 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.
- 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.

### Understand Code First
HARD-GATE: Do NOT write, plan, or fix until you READ existing code.
1. Search 3+ similar patterns (grep/glob) — cite file:line evidence.
2. Read existing files in target area — understand structure, base classes, conventions.
3. Run python .claude/scripts/code_graph trace <file> --direction both --json when .code-graph/graph.db exists.
4. Map dependencies via connections or callers_of — know what depends on your target.
5. Write investigation to .ai/workspace/analysis/ for non-trivial tasks (3+ files).
6. Re-read analysis file before implementing — never work from memory alone.
7. NEVER invent new patterns when existing ones work — match exactly or document deviation.
BLOCKED until: Read target files; Grep 3+ patterns; Graph trace (if graph.db exists); Assumptions verified with evidence.

## Reference Docs (READ before reviewing)
- docs/project-reference/code-review-rules.md
- {skill-specific reference docs — e.g., integration-test-reference.md for integration-test-review; backend-patterns-reference.md for backend reviews; frontend-patterns-reference.md for frontend reviews}

## Target Files
{explicit file list OR "run git diff to see uncommitted changes" OR "read all files under {plan-dir}"}

## Output
Write a structured report to plans/reports/{review-type}-round{N}-{date}.md with sections:
- Status: PASS | FAIL
- Issue Count: {number}
- Critical Issues (with file:line evidence)
- High Priority Issues (with file:line evidence)
- Medium / Low Issues
- Cross-cutting findings

Return the report path and status to the main agent.
Every finding MUST have file:line evidence. Speculation is forbidden.
`
})

Rules

• DO copy the template wholesale — including all 10 embedded protocol sections
• DO replace only the {placeholders} in Task / Round / Reference Docs / Target Files / Output sections with context-specific content
• DO choose code-reviewer subagent_type for code reviews and general-purpose for plan / doc / artifact reviews
• DO NOT paraphrase, summarize, or skip any protocol section
• DO NOT pass file contents inline — the sub-agent reads via its own tool calls so it has a fresh context
• DO NOT reference protocols by file path or tag name — the bodies are already embedded above
• DO NOT introduce placeholder markers for the protocols — they must stay literally expanded

• docs/project-reference/domain-entities-reference.md — Domain entity catalog, relationships, cross-service sync (read when task involves business entities/models) (content auto-injected by hook — check for [Injected:...] header before reading)
• docs/specs/ — Test specifications by module (read existing TCs; generate/update test specs via /tdd-spec after implementation)

Plan Quality — Every plan phase MUST ATTENTION include test specifications. 1. Add ## Test Specifications section with TC-{FEAT}-{NNN} IDs to every phase file 2. Map every functional requirement to ≥1 TC (or explicit TBD with rationale) 3. TC IDs follow TC-{FEATURE}-{NNN} format — reference by ID, never embed full content 4. Before any new workflow step: call TaskList and re-read the phase file 5. On context compaction: call TaskList FIRST — never create duplicate tasks 6. Verify TC satisfaction per phase before marking complete (evidence must be file:line, not TBD) Mode: TDD-first → reference existing TCs with Evidence: TBD. Implement-first → use TBD → /tdd-spec fills after.

Evidence Gate: MANDATORY IMPORTANT MUST ATTENTION — every claim, finding, and recommendation requires file:line proof or traced evidence with confidence percentage (>80% to act, <80% must verify first).

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 TaskCreate. 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. |

External Memory: For complex or lengthy work (research, analysis, scan, review), write intermediate findings and final results to a report file in plans/reports/ — prevents context loss and serves as deliverable.

Quick Summary

Goal: Implement a feature step-by-step with research, planning, execution, and verification.

Workflow:

1. Question — Clarify requirements via AskUserQuestion; challenge assumptions
2. Research — Use researcher subagents in parallel; scout codebase for patterns
3. Plan — Create implementation plan, get user approval
4. Implement — Execute with skill activation, code-simplifier, review-changes

Key Rules:

• Parent skill for all cook-* variants (cook-auto, cook-fast, cook-hard, cook-parallel)
• Write research findings to .ai/workspace/analysis/ for context preservation
• Always activate relevant skills from catalog during implementation
• Break work into small todo tasks; add final review task

Sub-Agent Selection — Full routing contract: .claude/skills/shared/sub-agent-selection-guide.md Rule: NEVER use code-reviewer for specialized domains (architecture, security, performance, DB, E2E, integration-test, git).

Frontend/UI Context (if applicable)

When this task involves frontend or UI changes,

UI System Context — For ANY task touching .ts, .html, .scss, or .css files: MUST ATTENTION READ before implementing: 1. docs/project-reference/frontend-patterns-reference.md — component base classes, stores, forms 2. docs/project-reference/scss-styling-guide.md — BEM methodology, SCSS variables, mixins, responsive 3. docs/project-reference/design-system/README.md — design tokens, component inventory, icons Reference docs/project-config.json for project-specific paths.

• Component patterns: docs/project-reference/frontend-patterns-reference.md
• Styling/BEM guide: docs/project-reference/scss-styling-guide.md
• Design system tokens: docs/project-reference/design-system/README.md

Pre-Implementation Granularity Gate (MANDATORY)

Plan Granularity — Every phase must pass 5-point check before implementation: 1. Lists exact file paths to modify (not generic "implement X") 2. No planning verbs (research, investigate, analyze, determine, figure out) 3. Steps ≤30min each, phase total ≤3h 4. ≤5 files per phase 5. No open decisions or TBDs in approach Failing phases → create sub-plan. Repeat until ALL leaf phases pass (max depth: 3). Self-question: "Can I start coding RIGHT NOW? If any step needs 'figuring out' → sub-plan it."

If ANY check fails → STOP. Ask user: "Phase needs more detail before implementation. Refine with /plan? [Y/n]" DO NOT implement a phase that contains planning verbs, unnamed files, or unresolved decisions.

Per-Phase Quality Cycle (MANDATORY)

Iterative Phase Quality — Score complexity BEFORE planning. Complexity signals: >5 files +2, cross-service +3, new pattern +2, DB migration +2 Score >=6 → MUST ATTENTION decompose into phases. Each phase: - ≤5 files modified - ≤3h effort - Follows cycle: plan → implement → review → fix → verify - Do NOT start Phase N+1 until Phase N passes VERIFY Phase success = all TCs pass + code-reviewer agent approves + no CRITICAL findings.

Each plan phase = one quality cycle (plan→implement→review→fix→verify). DO NOT start next phase until current phase passes VERIFY. After each phase: re-assess remaining phases for scope changes.

TC Satisfaction Verification (Per Phase)

After implementing each phase, before marking it complete:

1. Read the phase's ## Test Specifications section
2. For each mapped TC: verify evidence exists (file:line, not TBD), grep-verify the file
3. If any TC lacks evidence → phase is NOT complete
4. Update phase file's TC table with actual evidence references

Greenfield Mode

Auto-detected when no code directories (src/, app/, lib/, packages/) or manifests (package.json/*.sln/go.mod) exist.

1. Approved plan exists in plans/ → scaffold from plan
2. No plan → redirect: "Run /plan first to create a greenfield project plan."
3. Generate folder layout, starter files, build config, CI skeleton, CLAUDE.md
4. Skip codebase pattern search. After scaffolding, run /project-config.

Variant Decision Guide

Think harder to plan & start working on these tasks: $ARGUMENTS

Your Approach

• MUST ATTENTION use AskUserQuestion to clarify — NEVER assume requirements
• MUST ATTENTION be brutally honest — flag unrealistic/over-engineered approaches directly
• MUST ATTENTION present 2-3 alternatives with pros/cons for non-trivial decisions
• MUST ATTENTION challenge initial approach — the best solution often differs from first instinct

Workflow

IMPORTANT: Analyze the skills catalog at .claude/skills/* and activate needed skills during the process.

Research

• Parallel researcher subagents. Reports <=150 lines with citations.
• /scout-ext (preferred) or /scout (fallback) for codebase search.
• MUST ATTENTION write findings to .ai/workspace/analysis/{task-name}.analysis.md. Re-read ENTIRE file before planning.

Plan

• planner subagent with progressive disclosure: plan.md (<=80 lines) + phase-XX-name.md per phase.
• Each phase: Context, Overview, Requirements, Architecture, Related Files, Steps, TCs, Success Criteria, Risks, Next Steps.

Implementation

• /code to implement step by step. /interface-design for product UIs. /frontend-design for marketing/creative UIs.
• ui-ux-designer subagent for frontend per ./docs/design-guidelines.md.
• MUST ATTENTION run type checking and compile after each change.

Subagent Discipline: Paste full task text (NEVER make subagent read plan file). Require "ask questions before starting". Require self-review before reporting.

Batch Checkpoint (Large Plans)

For plans with 10+ tasks, execute in batches with human review:

1. Execute batch — Complete next 3 tasks (or user-specified batch size)
2. Report — Show what was implemented, verification output, any concerns
3. Wait — Say "Ready for feedback" and STOP. Do NOT continue automatically.
4. Apply feedback — Incorporate changes, then execute next batch
5. Repeat until all tasks complete

Testing

• Real tests: happy path, edge cases, error cases. NEVER fake data/mocks just to pass build.
• tester subagent → failures → debugger subagent → fix → repeat until green.
• MANDATORY: After writing tests, MUST run them and verify they pass. If tests fail, diagnose root cause: (a) test code has wrong setup/assertions → fix test, or (b) service code has actual bug → report as finding. Never mark test task as done until tests actually pass.

Code Review

Two-Stage Task Review — Both stages MUST ATTENTION complete before marking task done. Stage 1: Self-review — Immediately after implementation: - Requirements met? No regressions? Code quality acceptable? Stage 2: Cross-review — Via code-reviewer subagent: - Catches blind spots, convention drift, missed edge cases NEVER skip Stage 2. Self-review alone misses 40%+ of issues.

(1) spec-compliance-reviewer first, (2) code-reviewer after spec passes.

• Critical issues → fix → re-run tester. Report summary to user for approval.

PM, Docs & Final Report

• Approved: Parallel project-manager + docs-manager subagents. Rejected: Ask issues, fix, repeat.
• Final: summary of changes + next steps. Ask about commit/push via git-manager.

Red Flags — STOP

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

After implementing, run python.claude/scripts/code_graph connections <file> --json on modified files to verify no related files need updates.

Graph-Trace Before Implementation

MUST ATTENTION run BEFORE writing code when graph.db exists:

• python.claude/scripts/code_graph trace <file-to-modify> --direction both --json — callers + triggers
• python.claude/scripts/code_graph trace <file-to-modify> --direction downstream --json — all downstream consumers
• Prevents breaking implicit dependencies (bus consumers, event handlers) invisible in the file itself.

Workflow Recommendation

MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS: If you are NOT already in a workflow, you MUST ATTENTION use AskUserQuestion to ask the user. Do NOT judge task complexity or decide this is "simple enough to skip" — the user decides whether to use a workflow, not you: 1. Activate feature workflow (Recommended) — scout → investigate → plan → cook → review → sre-review → test → docs 2. Execute /cook directly — run this skill standalone

Next Steps (Standalone: MUST ATTENTION ask user via AskUserQuestion. Skip if inside workflow.)

MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS after completing this skill, you MUST ATTENTION use AskUserQuestion to present these options. Do NOT skip because the task seems "simple" or "obvious" — the user decides:

• "Proceed with full workflow (Recommended)" — I'll detect the best workflow to continue from here (feature implemented). This ensures review, testing, and docs steps aren't skipped.
• "/code-simplifier" — Simplify and clean up implementation
• "/integration-test" — Generate/update integration tests from test specs
• "/workflow-review-changes" — Review changes before commit
• "Skip, continue manually" — user decides

Standalone Review Gate (Non-Workflow Only)

MANDATORY IMPORTANT MUST ATTENTION: If this skill is called outside a workflow (standalone /cook), you MUST ATTENTION create a TaskCreate todo task for /review-changes as the last task in your task list. This ensures all changes are reviewed before commit even without a workflow enforcing it. If already running inside a workflow (e.g., feature, bugfix), skip this — the workflow sequence handles /review-changes at the appropriate step.

Closing Reminders

MANDATORY IMPORTANT MUST ATTENTION break work into small todo tasks using TaskCreate BEFORE starting. MANDATORY IMPORTANT MUST ATTENTION validate decisions with user via AskUserQuestion — never auto-decide. MANDATORY IMPORTANT MUST ATTENTION add a final review todo task to verify work quality. MANDATORY IMPORTANT MUST ATTENTION READ the following files before starting:

• MANDATORY IMPORTANT MUST ATTENTION verify all phases pass 5-point granularity check. Failing phases → sub-plan. "Can I start coding RIGHT NOW?" <!-- SYNC:understand-code-first:reminder -->
• MANDATORY IMPORTANT MUST ATTENTION search 3+ existing patterns and read code BEFORE any modification. Run graph trace when graph.db exists.
• MANDATORY IMPORTANT MUST ATTENTION include ## Test Specifications with TC IDs per phase. Call TaskList before creating new tasks.
• MANDATORY IMPORTANT MUST ATTENTION follow ALL steps regardless of perceived simplicity. "Too simple to plan" is an evasion, not a reason.
• MANDATORY IMPORTANT MUST ATTENTION read frontend-patterns-reference, scss-styling-guide, design-system/README before any UI change.
• MANDATORY IMPORTANT MUST ATTENTION score complexity first. Score >=6 → decompose. Each phase: plan → implement → review → fix → verify. No skipping.
• MANDATORY IMPORTANT MUST ATTENTION run at least ONE graph command on key files when graph.db exists. Pattern: grep → graph trace → grep verify.
• MUST ATTENTION apply critical thinking — every claim needs traced proof, confidence >80% to act. Anti-hallucination: never present guess as fact.
• MUST ATTENTION apply AI mistake prevention — holistic-first debugging, fix at responsible layer, surface ambiguity before coding, re-read files after compaction.

[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using TaskCreate.

[IMPORTANT] Analyze how big the task is and break it into many small todo tasks systematically before starting — this is very important.

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