Ralph Task Planning

Interview the user and create well-structured tasks for autonomous execution via tt ralph.

Interview Phase

Use AskUserQuestion to gather:

1. Goal: What are you trying to build or accomplish?
2. Done State: What does "done" look like? How do we verify?
3. Scope: What files/areas will this touch?
4. Constraints: Patterns to follow, things to avoid?

For codebase questions, use Task with subagent_type=Explore to research existing patterns.

Keep interview focused - 3-5 questions max.

Task Structure

Each task must have:

1. Clear completion criteria - How does Claude know it's done?
2. Verifiable success - Tests pass, typecheck passes, build works
3. Single responsibility - One focused outcome
4. Self-contained context - References specific files

Task Template

[Imperative description of what to do]

Context:
- Files: [specific paths to modify]
- Patterns: [reference existing code]

Success Criteria:
- [ ] [Specific outcome]
- [ ] Tests pass: pnpm test
- [ ] Types pass: pnpm typecheck

When complete, output: <promise>TASK_DONE</promise>

Good vs Bad Tasks

Good:

• "Add UserProfile type to src/types/user.ts with id, email, name fields"
• "Create getUserById in src/services/user.ts following existing service patterns"
• "Add unit tests for getUserById covering success and not-found cases"

Bad:

• "Implement user feature" (too vague)
• "Add types, service, and tests" (multiple things)
• "Make it work like the other module" (unclear reference)

Output

1. Present proposed task list with full prompts
2. Ask for approval/modifications via AskUserQuestion
3. Once approved, add tasks:

tt ralph --addTask "Task 1: [full task prompt]"
tt ralph --addTask "Task 2: [full task prompt]"

4. Show final list:

tt ralph --listTasks

5. Tell user to run tt ralph to start autonomous execution.