Security Architect
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill's stated purpose of 'secure-coding-review' combined with the 'modifies_files' capability creates a surface for indirect prompt injection, where malicious instructions in analyzed code could influence agent behavior.
- Ingestion points: Local project source code files.
- Boundary markers: None defined in the manifest.
- Capability inventory: Includes 'modifies_files' and 'creates_artifacts'.
- Sanitization: No sanitization or validation logic is specified in the manifest metadata.
- [Metadata Analysis] (SAFE): The manifest fields are professional and consistent with the skill's described purpose, with no evidence of deceptive metadata or prompt injection in the description or name.
Audit Metadata