The Agent Skills Directory

[SAFE]: The skill uses established industry tools (git and gh) to interact with GitHub, which is a well-known service. No unexpected or unauthorized network activity was detected.
[COMMAND_EXECUTION]: The skill executes standard repository management commands including git branch, git log, git diff, and git push, as well as GitHub CLI commands for issue and PR creation. These operations are essential for the skill's primary function.
[DATA_EXFILTRATION]: Data transfer is limited to pushing code and PR metadata to the user's configured GitHub remotes. This is a legitimate and expected behavior for a pull request automation tool.
[PROMPT_INJECTION]: The skill ingests untrusted data from git logs and diffs (ingestion points in SKILL.md) to generate summaries and titles for gh issue create and gh pr create (capability inventory). While it lacks boundary markers and explicit sanitization, the security risk is low as any potential injection would only affect the descriptive text of the resulting PR or issue.

pr