This skill helps AI agents prevent prompt injection by enabling real-time detection and collective immunity with open threat reporting and voting.
- Trust
- Needs review
- Install
- MCP ready
- Top risk
- Medium risk · Important install, trust, or metadata evidence is still thin.