Implements the RLM (Research-Lead-Manage) v3.0.0 methodology for structured software development with universal agent instructions, TDD, and prompt-based workflows across discovery, specification, implementation, and verification phases.
This skill has safety concerns that you should review before use. Some patterns were detected that may pose a risk.Safety score: 60/100.
KillerSkills scans all public content for safety. Use caution before installing or executing flagged content.
A comprehensive software development methodology that guides you through research, specification, implementation, and verification using structured prompts and documentation.
This skill implements the RLM (Research-Lead-Manage) v3.0.0 methodology, providing:
1. Always read these core files for context before any task:
- `AGENTS.md` — Universal agent instructions
- `RLM/START-HERE.md` — Workflow overview
- `RLM/specs/constitution.md` — Project standards and coding conventions
2. Configure auto-commits with semantic format:
- Format: `type(scope): description`
- Types: feat, fix, docs, refactor, test
- Reference RLM spec IDs when applicable (e.g., FTR-001, TASK-001)
When asked to "discover" or "research" an idea:
1. Read `RLM/prompts/01-DISCOVER.md` for the complete workflow
2. Check `RLM/research/project/` for existing research
3. Conduct 3-4 rounds of clarifying questions with the user
4. Generate PRD (Product Requirements Document) in `RLM/specs/`
5. Document research findings appropriately
When asked to create specs from a PRD:
1. Read `RLM/prompts/02-CREATE-SPECS.md`
2. Generate feature specifications
3. Create architecture documentation
4. Break down into epic-level components
5. Save all specs to `RLM/specs/`
When asked to create tasks:
1. Read `RLM/prompts/03-CREATE-TASKS.md`
2. Break features into 1-4 hour tasks
3. Ensure each task is atomic and testable
4. Save tasks to `RLM/tasks/active/`
5. Link tasks to parent feature specs
Follow this process for all implementation work:
**Step 1: Load Context (0-20%)**
**Step 2: Write Tests (20-40%)**
**Step 3: Implement (40-70%)**
**Step 4: Verify (70-85%)**
**Step 5: Review (85-100%)**
For design-related tasks:
1. Read `RLM/specs/PRD.md` for requirements
2. Create design system at `RLM/specs/design/design-system.md`
3. Define tokens at `RLM/specs/design/tokens.json`
4. For features: Create `design-spec.md` in feature directory
5. Ensure all components have 8 states:
- Default, Hover, Focus, Active, Disabled, Loading, Error, Empty
6. Follow WCAG 2.1 AA accessibility standards
When performing code review:
1. Check for security vulnerabilities (OWASP Top 10)
2. Identify performance issues (N+1 queries, memory leaks, etc.)
3. Verify code quality and standards compliance
4. For UI code: Check design token usage and accessibility
5. Create review report at `RLM/progress/reviews/`
6. Block commit if critical issues are found
For end-to-end verification:
1. Run Playwright tests for the feature
2. Execute functional, accessibility, and visual tests
3. On PASS: Update status to "verified"
4. On FAIL: Create bug tasks in `RLM/tasks/active/`
5. Generate verification report at `RLM/progress/verification/`
Throughout all phases:
- TypeScript/JavaScript: `npx eslint --fix`
- Python: `ruff check --fix`
```
User: "I want to add user authentication to the app"
1. Read RLM/prompts/01-DISCOVER.md
2. Ask clarifying questions about auth methods, providers, security requirements
3. Generate PRD.md with findings
4. Read RLM/prompts/02-CREATE-SPECS.md
5. Create feature spec, architecture docs, epic breakdown
6. Read RLM/prompts/03-CREATE-TASKS.md
7. Break into tasks like "Setup auth provider", "Create login UI", etc.
8. Save tasks to RLM/tasks/active/
```
```
User: "Implement task TASK-042: Create login form component"
Load Context (0-20%):
Write Tests (20-40%):
Implement (40-70%):
Verify (70-85%):
Review (85-100%):
```
```
User: "Review the authentication feature"
1. Check security: password hashing, session management, CSRF protection
2. Check performance: query optimization, caching strategy
3. Verify standards: follows constitution.md conventions
4. Check accessibility: WCAG 2.1 AA compliance for login UI
5. Create RLM/progress/reviews/auth-feature-review.md
6. If critical issues found: block commit and create bug tasks
```
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/rlm-method-workflow/raw