The Agent Skills Directory

[Command Execution]: The skill orchestrates multiple shell commands using the git and gh CLIs to automate the staging, committing, pushing, and creation of pull requests. This behavior is necessary for the skill's primary utility but requires appropriate monitoring of execution boundaries.
[Indirect Prompt Injection]: The skill ingests external content by discovery and reading repository pull request templates (e.g., .github/pull_request_template.md) and existing pull request bodies via gh pr view. If these files contain untrusted text or external contributions, there is a potential risk that the instructions embedded within them could influence the agent's subsequent behavior, as explicit parsing boundaries or sanitization steps are not defined.

yeet