5.8 KiB
name, description
| name | description |
|---|---|
| writing-plans | Use when design is complete and you need detailed implementation tasks for engineers with zero codebase context - creates comprehensive implementation plans with exact file paths, complete code examples, and verification steps assuming engineer has minimal domain knowledge |
Writing Plans
Overview
Write comprehensive implementation plans assuming the engineer has zero context for our codebase and questionable taste. Document everything they need to know: which files to touch for each task, code, testing, docs they might need to check, how to test it. Give them the whole plan as bite-sized tasks. DRY. YAGNI. TDD. Frequent commits.
Assume they are a skilled developer, but know almost nothing about our toolset or problem domain. Assume they don't know good test design very well.
Announce at start: "I'm using the writing-plans skill to create the implementation plan."
Save plans to: .claude/docs/plans/<feature-name>-plan.md
Testing Strategy: Auto-decide whether unit tests are needed based on complexity:
- Include tests for: Complex algorithms, business logic, data transformations where bugs are likely
- Skip tests for: Simple CRUD, UI components, straightforward mappings, anything you're 100% certain is bug-free
- Test type: Only deterministic unit tests - no integration tests, no complex mocking, no async complexity
- Tests should verify logic, not implementation details
Bite-Sized Task Granularity
Each step is one action (2-5 minutes):
If tests are included (for complex behavior):
- "Write the failing test" - step
- "Run it to make sure it fails" - step
- "Implement the minimal code to make the test pass" - step
- "Run the tests and make sure they pass" - step
- "Commit" - step
If no tests (for simple changes):
- "Implement the code" - step
- "Verify with typecheck/lint" - step
- "Commit" - step
Plan Document Header
Every plan MUST start with this header:
# [Feature Name] Implementation Plan
**Goal:** [One sentence describing what this builds]
**Architecture:** [2-3 sentences about approach]
**Tech Stack:** [Key technologies/libraries]
**Testing Approach:** [If tests included: "TDD with deterministic unit tests for complex logic (Tasks X, Y, Z)" | If no tests: "Code review and typecheck verification - logic is straightforward and bug-free"]
---
Task Structure
With Tests (Complex Behavior)
### Task N: [Component Name]
**Files:**
- Create: `exact/path/to/file.ts`
- Modify: `exact/path/to/existing.ts:123-145`
- Test: `tests/exact/path/to/test.spec.ts`
**Step 1: Write the failing test**
```typescript
describe('specific behavior', () => {
it('should return expected result', () => {
const result = function(input);
expect(result).toBe(expected);
});
});
```
**Step 2: Run test to verify it fails**
Run: `npm run test -- test.spec.ts`
Expected: FAIL with "function is not defined"
**Step 3: Write minimal implementation**
```typescript
export function function(input: InputType): ReturnType {
return expected;
}
```
**Step 4: Run test to verify it passes**
Run: `npm run test -- test.spec.ts`
Expected: PASS
**Step 5: Commit**
```bash
git add tests/path/test.spec.ts src/path/file.ts
git commit -m "feat: add specific feature"
```
Without Tests (Simple Changes)
### Task N: [Component Name]
**Files:**
- Create: `exact/path/to/file.ts`
- Modify: `exact/path/to/existing.ts:123-145`
**Step 1: Implement the code**
```typescript
export function function(input: InputType): ReturnType {
return expected;
}
```
**Step 2: Verify with typecheck**
Run: `npm run typecheck`
Expected: No errors
**Step 3: Verify with lint**
Run: `npm run lint`
Expected: No errors
**Step 4: Commit**
```bash
git add src/path/file.ts
git commit -m "feat: add specific feature"
```
Remember
- Auto-decide on unit tests - only for complex logic where bugs are likely
- Only deterministic unit tests - no integration/async/complex mocking
- Exact file paths always
- Complete code in plan (not "add validation")
- Exact commands with expected output
- Reference relevant skills with @ syntax
- DRY, YAGNI, frequent commits
- TypeScript syntax for all examples
Git Commit Guidance
IMPORTANT: Plans include commit steps, but executing-plans will handle commit approval.
During Execution (not plan writing):
- NEVER auto-commit - Even if plan includes commit commands, agent must ask user first
- When to suggest commits:
- After completing batch where agent verified tests pass
- Leverage existing pause points (batch completion, not extra pauses)
- Only when agent can verify the work (tests, typecheck, lint)
- When NOT to suggest commits:
- Tests are user's responsibility (agent can't verify)
- No natural pause point exists
- Manual verification required
In this skill (plan writing):
- Include commit steps in tasks as part of the plan
- Commit commands show what should be committed, not when
- executing-plans skill handles actual commit timing and user approval
Execution Handoff
After saving the plan, offer execution choice:
"Plan complete and saved to .claude/docs/plans/<filename>.md. Two execution options:
1. This Session (default) - Execute tasks iteratively in this session, reviewing and adjusting between tasks
2. Parallel Session - Open new session with executing-plans skill for batch execution with checkpoints
Which approach? (Press Enter for option 1)"
If This Session (default):
- Use the superpowers executing-plans skill in this session
- Execute in batches (default: first 3 tasks)
- Review and iterate between batches
- Adjust plan if needed based on learnings
If Parallel Session chosen:
- Guide them to open new session
- REQUIRED SUB-SKILL: New session uses superpowers executing-plans
- Load the plan file and execute in batches