Create detailed implementation strategies and architecture documentation
Installation
Details
Usage
After installing, this skill will be available to your AI coding assistant.
Verify installation:
npx agent-skills-cli listSkill Instructions
name: writing-plans category: workflow version: 1.0.0 description: Create detailed implementation strategies and architecture documentation author: obra/superpowers verified: true tags: [planning, architecture, documentation, strategy]
Writing Plans
Overview
Create detailed, actionable implementation plans that break down complex features into clear steps with proper architecture consideration.
When to Use This Skill
β Complex features requiring multiple steps β System design and architecture decisions β Technical specifications and implementation docs β Multi-person coordination
Plan Structure
Executive Summary
## Overview
[2-3 sentences explaining what we're building and why]
## Goals
- [ ] Primary goal: what success looks like
- [ ] Secondary goals: additional benefits
- [ ] Non-goals: explicitly out of scope
## Success Metrics
- Performance: [specific metrics]
- Quality: [test coverage, etc.]
- Timeline: [when it's needed]
Technical Approach
## Architecture
[High-level architecture description]
## Components
1. **Component A**: [purpose and responsibility]
2. **Component B**: [purpose and responsibility]
3. **Component C**: [purpose and responsibility]
## Data Flow
[How data moves through the system]
## Alternatives Considered
| Approach | Pros | Cons | Decision |
|----------|------|------|----------|
| Option A | Fast | Complex | β Not worth complexity |
| Option B | Simple | Slower | β
**Chosen** |
| Option C | Modern | Risky | β Too experimental |
Implementation Steps
## Phase 1: Foundation (Days 1-2)
- [ ] Step 1.1: [specific task]
- [ ] Step 1.2: [specific task]
- [ ] Step 1.3: [specific task]
**Dependencies**: None
**Deliverable**: [what's complete]
## Phase 2: Core Features (Days 3-5)
- [ ] Step 2.1: [specific task]
- [ ] Step 2.2: [specific task]
- [ ] Step 2.3: [specific task]
**Dependencies**: Phase 1 complete
**Deliverable**: [what's complete]
## Phase 3: Integration & Polish (Days 6-7)
- [ ] Step 3.1: [specific task]
- [ ] Step 3.2: [specific task]
- [ ] Step 3.3: [specific task]
**Dependencies**: Phase 2 complete
**Deliverable**: [production-ready feature]
Testing Strategy
## Unit Tests
- [ ] Core business logic
- [ ] Edge cases and error handling
- [ ] Data validation
## Integration Tests
- [ ] API endpoints
- [ ] Database operations
- [ ] External service integrations
## Manual Testing
- [ ] User flows
- [ ] Performance under load
- [ ] Cross-browser/device testing
## Test Coverage Goal: >80%
Risk Assessment
## Technical Risks
| Risk | Impact | Probability | Mitigation |
|------|--------|-------------|------------|
| Performance bottleneck | High | Medium | Implement caching, add monitoring |
| Third-party API changes | Medium | Low | Version pinning, fallback strategy |
| Data migration complexity | High | High | Thorough testing, rollback plan |
## Dependencies
- [ ] External API approval
- [ ] Database schema approval
- [ ] Security review
## Rollback Plan
[How to revert if things go wrong]
Planning Best Practices
Make Steps Atomic
β Bad: "Build the payment system" β Good: "Create Stripe checkout form"
Include Definition of Done
Each step should have clear completion criteria:
- Code written and reviewed
- Tests passing
- Documentation updated
- No regressions
Consider Edge Cases
## Edge Cases to Handle
- What happens if the API is down?
- How do we handle malformed input?
- What if the user has no internet connection?
- Concurrent requests for the same resource?
- Extremely large datasets?
Plan for Iteration
## Iteration 1 (MVP)
- Core functionality
- Basic error handling
- Happy path only
## Iteration 2
- Enhanced error handling
- Edge case coverage
- Performance optimization
## Iteration 3
- Advanced features
- Polish and UX improvements
Common Planning Mistakes
β Skipping the "why" - context matters β No alternative analysis - first idea isn't always best β Unrealistic timelines - buffer time for the unknown β Ignoring risks - what could go wrong? β Testing as afterthought - plan tests upfront β No rollback plan - assume you'll need it
Integration with Claude
When planning, say:
- "Help me create a plan for [feature]"
- "Break down this implementation into steps"
- "What are the risks with this approach?"
- "Create an architecture doc for [system]"
Claude will:
- Structure comprehensive plans
- Identify potential issues
- Suggest alternatives
- Break down complex tasks
- Ensure nothing is forgotten
- Create clear, actionable steps
More by Lordsisodia
View allFour-phase root cause analysis process to find bugs 10x faster
Build custom Model Context Protocol servers to extend Claude's capabilities
Extract text, tables, metadata from PDFs with merge and annotation support
Build complex Claude.ai HTML artifacts using React, Tailwind CSS, and shadcn/ui
