The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill manages the installation and execution of a GitHub CLI extension (gh-attach). It references a vendor-owned version (MTGVim/gh-attach) and a third-party version (enthus-appdev/gh-attach). Although it requires explicit user confirmation before installing unreviewed versions, the mechanism involves executing remote code from a third-party repository.
[INDIRECT_PROMPT_INJECTION]: The skill processes existing GitHub PR body content, which is externally controlled and could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill reads the existing PR body from the GitHub API as described in SKILL.md (Workflow Step 1).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content within the ingested PR body.
Capability inventory: The skill can execute shell commands (gh attach), perform network operations via the GitHub API, and utilize browser automation (Playwright).
Sanitization: No input sanitization is performed on the ingested PR body, though the skill does include instructions to redact signed parameters from output URLs.
[COMMAND_EXECUTION]: The skill executes multiple shell commands on the host system, including gh extension list, gh extension install, gh attach, git check-ignore, and git ls-files.
[EXTERNAL_DOWNLOADS]: The skill identifies and potentially downloads code from external sources, specifically github.com/enthus-appdev/gh-attach and github.com/MTGVim/gh-attach.

tk-github-image-upload-to-pr