Security First
Our security scoring system helps you identify safe, trustworthy AgentSkills for your OpenClaw deployment.
Understanding Security Scores
High Security (80-100)
Excellent security practices, minimal permissions, regular updates, and strong community trust.
Medium Security (60-79)
Generally safe but may have some concerns like broad permissions or limited updates.
Low Security (0-59)
Significant security concerns. Use with caution and only in isolated environments.
How We Score Security
Code Quality
Permissions & Access
Data Handling
Community & Maintenance
Top Skills by Security Score
Live data from ClawHub — updated every 5 minutes
Security Best Practices
Check Permissions
Always review the permissions a skill requests. More permissions = higher risk.
Review Code Quality
Look for skills with good documentation, error handling, and type safety.
Check Update History
Regularly updated skills are more likely to have security fixes.
Read Community Reviews
Community feedback often reveals security concerns not caught in audits.
Want to Audit Your Skill?
Submit your AgentSkill for a free security audit. Get detailed feedback and improve your security score.
Average audit time: 3-5 business days • Free for open-source skills