The Agent Skills Directory

[SAFE]: The skill provides meta-instructions for the agent to author and organize repository-based skills. It includes a Python snippet for validation that correctly uses yaml.safe_load to prevent unsafe deserialization.
[SAFE]: While the skill mentions file system paths (e.g., /home/bb/hermes-agent/skills/ and ~/.hermes/skills/), these are used to define the organizational structure of the agent's own skills and do not involve unauthorized access or exfiltration of sensitive data.
[SAFE]: No remote code execution, obfuscation, prompt injection, or persistence mechanisms were detected. The skill focuses on standard software development workflows like file creation and git operations within its designated scope.

hermes-agent-skill-authoring