excavate
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs extensive analysis using local shell commands,
git, and the GitHub CLI (gh). These tools are used to extract revision history, PR bodies, and review conversations to build a knowledge map of the repository. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git logs and GitHub PR/issue descriptions. To mitigate risks, it follows a 'Grounding Rule' where the model only writes prose around mechanically-derived data and implements an automated redaction system to mask potential secrets and personal identifiers before processing text.
- [DYNAMIC_EXECUTION]: The skill manages its analysis pipeline by generating and sourcing shell scripts within a unique temporary directory. This directory is created with restricted permissions (
mode 700) to prevent local data exposure during the mining process. - [SAFE]: The skill demonstrates a strong security posture by incorporating built-in redaction for various secret patterns (e.g., AWS keys, GitHub tokens, Slack tokens), contributor name scrubbing, and requiring human confirmation for every generated entry before committing it to the codebase.
Audit Metadata