メニュー
Use this skill to break work into small, verifiable chunks. Activates when planning implementation, breaking down features, or when tasks seem too large.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | micro-tasking |
| description | Use this skill to break work into small, verifiable chunks. Activates when planning implementation, breaking down features, or when tasks seem too large. |
Break work into 2-5 minute tasks with clear completion criteria.
Large tasks fail. Small tasks succeed.
A task that takes more than 10 minutes is not a task—it's a project that needs decomposition.
Every micro-task has:
task:
id: TASK-001
title: Short, action-oriented description
duration: 2-5 minutes
file: exact/path/to/file.ts
line_hint: ~50-75 # approximate location
action: |
Specific action to take
verification: |
How to know it's done
depends_on: [] # or [TASK-000]
feature: "Add user login"
tasks:
- Create LoginForm component skeleton (3 min)
- Add email input with label (3 min)
- Add password input with label (3 min)
- Add submit button (2 min)
- Add form onSubmit handler (3 min)
- Add loading state to button (3 min)
- Add error message display (3 min)
- Connect to auth API (5 min)
- Add success redirect (3 min)
- Write unit test for form render (5 min)
- Write unit test for submission (5 min)
bug: "Login fails silently"
tasks:
- Add console.log to trace request (2 min)
- Verify API endpoint is correct (2 min)
- Check response handling (3 min)
- Add error state to component (3 min)
- Display error message to user (3 min)
- Add test for error case (5 min)
refactor: "Extract shared button styles"
tasks:
- Identify all button variants (3 min)
- Create Button component file (2 min)
- Add base button styles (3 min)
- Add variant prop (primary/secondary) (3 min)
- Replace first usage (3 min)
- Replace second usage (3 min)
- Verify styles match original (2 min)
- Run visual regression test (3 min)
- id: TASK-XXX
title: Create [Component] skeleton
duration: 3 min
file: src/components/[Component].tsx
action: Create component with props interface, return null
verification: Component imports without error
- id: TASK-XXX
title: Add test for [behavior]
duration: 5 min
file: src/components/[Component].test.tsx
action: Write describe block with it() for [behavior]
verification: Test runs (can be failing)
- id: TASK-XXX
title: Add error handling for [case]
duration: 3 min
file: src/lib/[file].ts
line_hint: ~45
action: Add try/catch around [operation]
verification: Error case shows message instead of crashing
For each task:
1. Announce: "Starting TASK-XXX: [title]"
2. Set timer: 5 minutes max
3. Do the ONE thing
4. Verify completion criteria
5. Commit if appropriate
6. Announce: "Completed TASK-XXX"
7. If over time: Task was too big, note for future
| Aspect | Large Tasks | Micro-Tasks |
|---|---|---|
| Progress visibility | Unclear | Clear |
| Context switching | Loses state | Easy resume |
| Estimation accuracy | Way off | Reasonable |
| Verification | Hard | Easy |
| Morale | Frustrating | Satisfying |
Works with:
/ledger - Track micro-task completion/execute-plan - Execute task by task/riper - Plan phase creates micro-tasksWHEN to auto-invoke: Creating API routes, building endpoints, adding route.ts files, implementing REST/GraphQL APIs, adding authentication to APIs, rate limiting, API validation with Zod, handling HTTP methods (GET/POST/PUT/DELETE). WHEN NOT to invoke: Pure frontend work, styling, React components without API calls, documentation-only tasks, database migrations without API changes.
WHEN to auto-invoke: Code review requests, refactoring code, cleaning up code, performance optimization, security audits, linting issues, reducing technical debt. WHEN NOT to invoke: Building new features (use domain-specific skills), initial implementation, documentation-only work.
WHEN to auto-invoke: End of productive sessions, after solving difficult problems, after user corrections, after discovering effective workarounds, after establishing project conventions. WHEN NOT to invoke: Trivial sessions, sessions with only exploration, when explicitly told not to learn.
WHEN to auto-invoke: Database schema design, creating migrations, writing SQL queries, query optimization, Supabase operations, Prisma/Drizzle schema changes, PostgreSQL tasks, RLS policies, indexes. WHEN NOT to invoke: Pure frontend work, API logic without database, styling, React components, deployment tasks.
WHEN to auto-invoke: Creating CI/CD pipelines, GitHub Actions workflows, Docker/containerization, Vercel/Netlify deployment, infrastructure as code, environment variables, deployment configurations. WHEN NOT to invoke: Application code changes, UI development, database schema design, testing application logic.
WHEN to auto-invoke: Measuring code quality, testing AI outputs, establishing success criteria, defining acceptance tests, checking implementation correctness. WHEN NOT to invoke: Simple implementations, prototyping, exploration phases, when speed matters more than measurement.