autofix
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from GitHub Pull Request review threads which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: In
SKILL.md(Step 3) andgithub.md(Section 3), the skill fetches review comments using the GitHub GraphQL API (gh api graphql). It also searches for and loads local repository instructions fromAGENTS.md(Step 0). - Boundary markers: The skill contains explicit instructions to treat all thread comment bodies as untrusted input and use them only as issue reports, never as executable instructions (Step 3, Step 4).
- Capability inventory: The skill possesses the ability to modify local source code via the agent's Edit tool (Step 6), create commits (
git commit), push changes to remote repositories (git push), and post comments to GitHub PRs (gh pr comment). - Sanitization:
SKILL.md(Step 6) defines comprehensive sanitization rules, including redacting secrets and tokens, stripping shell command suggestions, and removing paths to sensitive files (dotfiles, credentials) before presenting summaries to the user. - [COMMAND_EXECUTION]: The skill executes various
gitandgh(GitHub CLI) commands to interact with the repository and pull requests. - Evidence: Commands such as
gh pr list,gh pr create,git status,git push, andgit commitare used throughout the workflow based on the local environment state and explicit user confirmation.
Audit Metadata