Community Leaderboard
Contributors who make the AI agent ecosystem safer for everyone. Every rule proposed and promoted protects thousands of developers.
Loading contributors...
How to contribute
- Install Guard:
npm i -g @panguard-ai/panguard-guard - Scan your skills:
panguard-guard scan - Start Guard:
panguard-guard start - Guard automatically proposes rules from threats it detects
- Rules that pass community review protect everyone
Propose
Turn a threat into a rule
Guard surfaces suspicious skill behavior and drafts a detection rule from the evidence, so contributing starts from real signal, not guesswork.
Review
Community + automated review
Each proposal is checked against a benign corpus for false positives before it can be promoted, keeping the shared ruleset precise.
Protect
Everyone gets safer
Promoted rules ship to every Guard install. One contributor's catch protects thousands of downstream agents.
Contribute on GitHub
The Agent Threat Rules standard is open source and MIT-licensed. Browse rules, open issues, and submit pull requests.