docs-review

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFE
Full Analysis
  • [Command Execution]: The skill utilizes the gh command-line interface to perform actions like viewing pull requests (gh pr view), reading diffs (gh pr diff), and posting review comments (gh api). This is a standard practice for automation skills designed to integrate with GitHub workflows.- [Data Ingestion and Processing]: The agent is instructed to read external content from pull request diffs and documentation files (src/content/docs/*). This creates a potential surface for indirect prompt injection where instructions could be embedded in the documentation being reviewed. However, the skill's specific focus on MDX syntax and style guides provides a clear scope that mitigates the likelihood of the agent being diverted from its primary task.- [File System and Repository Interaction]: Based on the decision logic, the agent may edit local documentation files and commit changes directly. This capability is restricted to the repository context and is intended for fixing obvious technical errors like broken code or syntax issues in documentation, which is a common developer workflow requirement.- [Network Operations]: The skill performs network requests specifically to GitHub's REST API via the gh tool. These operations are targeted toward official GitHub endpoints for pull request management, which is an expected behavior for a review tool.
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 06:46 AM