The Agent Skills Directory

[COMMAND_EXECUTION]: Includes template bash commands using curl and jq to interact with the GitHub API. These commands are used to retrieve repository statistics and list file contents for the purpose of security auditing.
[EXTERNAL_DOWNLOADS]: Fetches metadata and skill configuration files from official GitHub domains (api.github.com and raw.githubusercontent.com). These are well-known services and the retrieval is limited to data for analysis.
[PROMPT_INJECTION]: The skill is intended to process untrusted data from external repositories as part of its vetting protocol. This represents an indirect prompt injection surface; however, the skill provides a structured framework for evaluation and reporting to help the agent maintain a critical security posture.

skill-vetter