agent-security-review
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE] (SAFE): No malicious code or behavioral patterns were detected in the skill. All scripts and descriptions focus on promoting secure coding practices such as secrets management, input validation, and SQL injection prevention.
- [Indirect Prompt Injection] (SAFE): The skill acts as an advisor for auditing untrusted code, which is a known ingestion surface. However, the skill itself lacks active capabilities or scripts that could be exploited. 1. Ingestion points: User-provided source code files for analysis. 2. Boundary markers: None explicitly defined. 3. Capability inventory: No executable scripts or tool-based capabilities. 4. Sanitization: Not applicable as the skill is instructional documentation.
Audit Metadata