Scan every skill
before it gets airborne.

Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks.

THREAT READOUTSES//04A · LIVE

7findings across 142 skills scanned

▲shell.exec → unescaped argvHIGH

◆fetch() to undeclared hostMED

◆secret-like literal in promptMED

●manifest signature verifiedPASS

Readout values are illustrative.

SASTstatic skill analysis

Patternsmalicious-intent heuristics

CIgate on every commit

OSSopen rule registry

Scan every skillbefore it gets airborne.