The Agent Skills Directory

Command Execution via GitHub CLI: The skill utilizes the gh (GitHub CLI) tool to perform repository operations, including viewing PR details, checking CI status, and posting comments. These commands are executed within the scope of pull request maintenance.
Ingestion of Untrusted External Data: The skill is designed to read pull request titles, descriptions, and review comments from the docker/docs repository.
Ingestion points: Data enters the agent's context through gh pr view and gh api calls that retrieve user-generated comments and PR metadata (SKILL.md).
Boundary markers: There are no explicit instructions or delimiters to isolate or ignore potentially malicious instructions embedded within these external comments.
Capability inventory: The agent has the ability to execute shell commands, modify repository content, commit changes, and interact with the GitHub API.
Sanitization: The instructions do not specify any validation, filtering, or escaping of the content retrieved from GitHub before the agent acts upon it.
Repository Interaction and Modification: The skill includes instructions for the agent to commit and push changes back to the repository branch. While this is the primary purpose of the skill, it allows the agent to introduce code changes based on its interpretation of review comments.
Disclosure Practices: The instructions include a requirement for the agent to provide a disclosure footer (e.g., 'Generated by Codex') when replying to comments, which is a helpful transparency measure for automated interactions.

maintain-pr