Free Tool
Is this AI skill safe?
Paste any SKILL.md file and we'll scan it for prompt injection, data exfiltration, credential theft, and malicious code. 36% of public AI skills contain security issues.
Drop a file or click to upload
.md.txt.zipMax 500KB
76 pattern rules + AI deep scan
What we scan for
Prompt injection & jailbreaks
Data exfiltration to external servers
Credential theft & API key leaks
Malicious code execution (eval, exec, curl|bash)
AI model token injection & role hijacking
Obfuscation (base64, Unicode, zero-width chars)
Supply chain attacks (remote packages, scripts)
Social engineering & fake error messages
76 pattern rules+AI deep scan powered by Claude