create-pr

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs expected repository management tasks using standard development tools like git and gh.
  • [PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface when processing external data.
  • Ingestion points: Commit messages via git log, code changes via git diff, and repository-hosted pull request templates in locations like .github/pull_request_template.md.
  • Boundary markers: None explicitly defined in the instructions to protect against malicious instructions embedded in the templates or commits.
  • Capability inventory: Uses 'gh pr create' to send data to GitHub and 'git push' to upload code.
  • Sanitization: No specific sanitization or filtering of commit messages or template content is implemented before interpolation into the PR creation command.
  • [COMMAND_EXECUTION]: Executes shell commands to interact with the git environment and the GitHub API. The tools used (git, gh, cat, ls, grep, sed) are appropriate for the skill's stated purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 08:26 AM
Security Audit — agent-trust-hub — create-pr