The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of pull request review comments and general comments from the GitHub API.
Ingestion points: In SKILL.md, the /git pr-pull-review command fetches data via gh pr view and gh api from potentially attacker-controlled sources (PR reviewers).
Boundary markers: The skill requires the agent to present a summary of comments and ask the user which to address before taking action, providing a human-in-the-loop check.
Capability inventory: The agent has the ability to read and edit project files, execute git commands, and interact with the GitHub API (replying to comments).
Sanitization: There is no explicit mention of sanitizing or escaping the content of comments before they are processed by the agent.
[COMMAND_EXECUTION]: The skill relies on executing various shell commands to perform its core functions.
Evidence: Extensive use of git (status, diff, add, commit, log, push) and gh (pr create, pr view, pr edit, api) commands to manage the development workflow.

git