The Agent Skills Directory

[SAFE]: The skill performs legitimate operations to finalize task workflows, such as reading review records, drafting pull request descriptions, and managing follow-up tasks.
[SAFE]: No network activity was detected; the skill explicitly avoids remote operations and only interacts with the local file system and Git worktrees.
[SAFE]: All command executions are restricted to internal plugin scripts and standard system utilities like Git and jq, with proper argument quoting to prevent command injection.

completion