sec-threat-model
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill explicitly prohibits unauthorized network access, port scans, and access to third-party systems, ensuring all analysis remains local and authorized. It also restricts file writes to a specific temporary scan directory to prevent accidental source code modification.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted source code, creating a potential surface for indirect prompt injection. This is mitigated by instructions to prioritize evidence found in the code over claims in README files and to mark unverified data as 'open'. Ingestion points: source code files. Boundary markers: absent. Capability inventory: file read (source) and file write (output directory). Sanitization: not applicable as the skill analyzes code without execution.
Audit Metadata