Every skill analyzed every 6 hours
Secure Skills Hub
Browse 22951+ OpenClaw skills with independent security analysis. Every skill scanned. No AI. No black boxes.
22951
Skills Scanned
0
Flagged Risky
11
Security Signals
How Clawkeeper Scores Skills
Deterministic. No AI. No black boxes. Every skill is scored using 11 weighted signals that detect supply chain risks, suspicious code, and packaging anomalies.
| Signal | Weight | Severity | Description |
|---|---|---|---|
| clawhub_flagged | +3.0 | critical | Flagged by ClawHub moderation system |
| suspicious_handler | +2.0 | high | Contains executable handler code |
| executable_content | +2.0 | high | Files with executable MIME types |
| hidden_dotfile | +1.5 | high | Hidden dotfiles (.env, .ssh, etc.) |
| binary_file | +1.5 | high | Binary files in skill package |
| new_publisher_many_files | +1.5 | medium | New publisher with complex package |
| suspicious_filename | +1.5 | medium | Filenames containing backdoor/exploit/payload |
| missing_skill_md | +1.0 | low | No skill.md documentation when files exist |
| large_file | +1.0 | low | Files larger than 100KB (anomalous for skills) |
| rapid_version_churn | +1.0 | medium | Rapid version publishing pattern |
| excessive_files | +0.5 | info | Unusually high file count for a skill |
Verdict Thresholds
Clean
0 – 1.0
Low
1.1 – 3.0
Medium
3.1 – 5.0
High
5.1 – 7.0
Critical
7.1 – 10.0
Risk score = sum of all triggered signal weights. Capped at 10.0.
Continuous Scanning
Every skill in the ClawHub registry is rescanned every 6 hours. Verdicts update automatically.
Clean (1000)
Last scan: Jun 30, 2026, 12:00 AM
Protected by Clawkeeper
Deploy your first secured agent
Free to scan. Pro to deploy. Start protecting your AI agents today.