| name | multi-agent-validator |
| description | Coordinated validation protocol where multiple agent perspectives verify a solution. |
| category | meta |
| version | 1 |
Multi-Agent Validator
Purpose
Use multiple agent viewpoints to validate critical decisions and catch blind spots.
Validation Patterns
1. Adversarial Review
- ARCHITECT challenges ENGINEER's implementation
- SECURITY finds vulnerabilities in logic
- DEBUGGER identifies edge case failures
2. Consensus Building
- Each agent rates confidence 1-10
- Disagreements trigger debate protocol
- Final verdict requires 4+ votes
3. Cross-Domain Check
- Messenger expert reviews navigation code
- Performance expert reviews business logic
- Security expert reviews UX decisions
Output Format
## MULTI-AGENT VALIDATION
Architect: ✅ (confidence: 8)
Engineer: ✅ (confidence: 9)
Security: ⚠️ (needs auth check)
Debugger: ✅ (confidence: 7)
**Result: 3/4 ready, minor revision needed**