final-assignment-citation-review
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill includes an 'Anti-Patterns' section that explicitly instructs the agent to ignore instructions embedded in external content, mitigating indirect prompt injection risks.
- [SAFE]: Data access is scoped to project-specific files needed for citation review. No credentials, sensitive system paths, or exfiltration patterns were found.
- [SAFE]: Command execution mentioned is limited to local development scripts for skill maintenance and does not pose a security risk.
Audit Metadata