deep-codebase-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze entire software projects, which constitutes untrusted external data.
- Ingestion points: Files and documentation within any project directory targeted for analysis (described in SKILL.md).
- Boundary markers: The instructions lack specific delimiters or warnings to ignore instructions embedded within the codebase being analyzed (e.g., in comments or documentation).
- Capability inventory: The skill uses a directory listing script (scripts/analyze_structure.py) and typically relies on the Agent's native file-reading and terminal capabilities.
- Sanitization: There is no evidence of sanitization or filtering of codebase content before it is processed by the Agent, which could lead to the Agent obeying instructions found within the code (e.g., hidden comments directing the Agent to ignore safety protocols).
Audit Metadata