ai-security-guide
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill serves as a static documentation resource providing informational content about AI security. No malicious behavior, network exfiltration, or unauthorized file access was detected.
- [PROMPT_INJECTION]: Static analysis flagged potential prompt injection patterns (e.g., 'Ignore previous instructions'). Manual review confirms these are benign educational examples used to describe attack types in a security guide, not functional instructions intended to override agent behavior.
Audit Metadata