The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to invoke the GitHub CLI (gh) for fetching PR status, check logs, and review comments. This is the intended primary purpose of the skill and is performed using standard programmatic patterns.
[COMMAND_EXECUTION]: The skill executes git and gh commands for committing changes and watching PR checks as part of the automated iteration workflow defined in SKILL.md.
[SAFE]: The scripts use uv run for execution, ensuring a reproducible Python environment. The logic includes specific classification for 'review bots' (e.g., Sentry, CodeQL) to ensure security and quality findings are not silenced, while informational bots (e.g., Codecov) are correctly ignored.

iterate-pr