The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) to interact with pull requests and issues. These operations are within the scope of the skill's intended purpose and are restricted by the allowed-tools frontmatter configuration.
[INDIRECT_PROMPT_INJECTION]: The skill contains a verification step that reads external data from GitHub via gh pr view --json body,comments. This represents a potential injection surface where an attacker could place malicious instructions in a PR comment. However, the risk is negligible as the skill only processes this data to count occurrences of 'user-attachments' using grep -c, and does not execute instructions contained within the content. Sanitization is implicitly handled by the grep count logic.
[SAFE]: All network operations and tool interactions are directed towards GitHub, a well-known and trusted service. The skill follows best practices by instructing the user to handle upgrades and authentication manually rather than attempting to automate these sensitive tasks.

github-image-upload