meta-prompt

Meta-Prompt Command

Let Claude refine complex objectives into concrete, executable workflows with governance validation.

Philosophy

"Tell Claude what you want, not what to do"

Instead of manually designing agent sequences, describe your high-level goal. Meta-prompting:

• Refines your objective into specific requirements
• Proposes optimal agent workflows
• Validates against governance rules
• Provides alternatives and trade-offs
• Explains reasoning behind suggestions

---

When to Use

Perfect for:

• Complex, multi-step projects
• Unfamiliar domains (don't know which agents to use)
• Architectural decisions (multiple valid approaches)
• Exploring solution space (what's possible?)
• Learning optimal patterns

Not needed for:

• Simple, single-agent tasks
• Well-known patterns you've done before
• When you already know exact agent sequence

---

Basic Usage

Describe what you want to achieve:

/meta-prompt "I need to build a user authentication system with email verification"

What happens:

1. Analyzes your objective
2. Identifies requirements and constraints
3. Proposes concrete workflow with agents
4. Validates against governance (agents exist, budget, skills)
5. Provides alternatives and trade-offs
6. Explains reasoning

Output:

## Meta-Prompting Analysis

### Refined Objective
Build a secure user authentication system with:
- Email/password registration
- Email verification (confirm email before access)
- Secure password storage (bcrypt/argon2)
- JWT-based session management
- Password reset flow
- Rate limiting for security

### Proposed Workflow

**Recommended**: `full-stack-feature` (customized)

**Agent Sequence** (8 agents, ~45-60 min, $3.20-4.50):
1. **backend-architect** - Design auth API endpoints, JWT strategy
2. **database-architect** - User schema, indexes, migrations
3. **security-specialist** - Review password storage, rate limiting, vulnerabilities
4. **python-expert** - Implement auth endpoints with FastAPI/Flask
5. **frontend-developer** - Login/register UI components
6. **email-specialist** - Email verification flow (optional: use existing templates)
7. **qc-automation-expert** - Security tests (SQL injection, XSS, CSRF)
8. **deployment-integration-expert** - Environment variables, secrets management

### Reasoning

**Why this sequence:**
- Backend first ensures API contract for frontend
- Security specialist early to catch design issues (cheaper than fixing later)
- Email specialist optional if using third-party service (Sendgrid, etc.)
- QA must include security tests for auth systems

**Alternatives Considered:**
1. **Simpler**: Skip email verification ($2.50, 35min)
   - Trade-off: Less secure, but faster to market
2. **More robust**: Add 2FA, OAuth providers ($6.00, 90min)
   - Trade-off: Better security, more complex

### Governance Validation

✅ **All checks passed**

- ✅ All agents available
- ✅ Required skills present (backend, database, security, frontend)
- ✅ Budget: $4.50 (within typical range)
- ✅ No security policy violations
- ✅ Timeline: ~60 min (reasonable for scope)

### Success Criteria

I'll know this is done when:
- [ ] Users can register with email/password
- [ ] Email verification required before access
- [ ] Passwords securely hashed (bcrypt/argon2)
- [ ] JWT tokens issued and validated
- [ ] Password reset flow works
- [ ] Security tests pass (no SQL injection, XSS, CSRF)
- [ ] Rate limiting prevents brute force

### Risk Assessment

**Potential risks:**
- ⚠️ Email delivery (use reliable service, test thoroughly)
- ⚠️ Token expiration handling (document clearly)
- ⚠️ Password reset security (verify email ownership)

---

**Ready to proceed?**
  [1] Run recommended workflow
  [2] Run alternative 1 (simpler)
  [3] Run alternative 2 (more robust)
  [4] Customize workflow
  [5] Save for later (add to todos)
  [0] Cancel

Your choice: