review-code

Pass

Audited by Gen Agent Trust Hub on Jul 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes several command-line tools to perform its tasks, including git for repository management, gh (GitHub CLI) for PR interactions, and ruff for linting. These are standard, well-known development tools used appropriately for the skill's stated purpose.
  • [EXTERNAL_DOWNLOADS]: The skill fetches data from GitHub's official API/CLI, including pull request metadata, diffs, and file contents. These network operations are directed at a well-known service (GitHub) and are necessary for retrieving the code to be reviewed.
  • [PROMPT_INJECTION]: By design, the skill processes external content (PR descriptions, commit messages, and source code) which presents a surface for indirect prompt injection. However, this is inherent to the nature of a code review tool, and the skill does not use the ingested data in dangerous execution contexts.
  • [SAFE]: The skill implements safety-conscious workflows, such as using git worktree in /tmp/ to avoid polluting the user's active working directory and checking for PR states (draft, closed, merged) before proceeding.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 5, 2026, 08:16 PM
Security Audit — agent-trust-hub — review-code