codebase-deep-research
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill leverages system commands such as git and ripgrep (rg) to perform repository discovery and symbol searching. It also allows for the execution of repro scripts to validate research findings.\n- [PROMPT_INJECTION]: The skill processes untrusted local codebase files (ingestion points) and possesses capabilities such as spawning subagents and executing generated repro scripts. While this creates a surface for indirect prompt injection, the skill includes boundary markers (defining research scope and targets) and instructs the agent to verify evidence against primary source files (sanitization) to mitigate risks.
Audit Metadata