The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses python3 -c to parse JSON responses from the GitHub API. The Python code is a static string included in the skill, and the piped input from curl is processed as data (JSON), not as executable code. This is a standard and safe practice when JSON-specific tools like jq are not available.
[DATA_EXPOSURE]: The skill provides instructions for managing GitHub Secrets, including a command to upload the user's private SSH key (~/.ssh/id_rsa). While this involves accessing sensitive data, it is the documented and intended purpose of the 'Secrets Management' section. No evidence of unauthorized or hidden exfiltration was found.
[CREDENTIALS_UNSAFE]: The skill manages authentication by checking for the gh CLI or a GITHUB_TOKEN in the environment. It also provides a fallback to read the token from a local .env file (~/.hermes/.env). This follows standard development practices for secure credential management and avoids hardcoding sensitive information.

github-repo-management