code-deduplication
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a process for the agent to catalog existing code capabilities and check them before writing new code. This is a common and safe practice for code maintenance.\n- [PROMPT_INJECTION]: The skill involves processing existing codebase data (headers, docstrings), which presents a standard surface for indirect prompt injection common to all code-analysis agents.\n
- Ingestion points: Source code files and the generated CODE_INDEX.md file.\n
- Boundary markers: None provided in the instructions for data read from files.\n
- Capability inventory: File reading, codebase searching (grep), and file writing (CODE_INDEX.md).\n
- Sanitization: None.
Audit Metadata