The Agent Skills Directory

[SAFE]: The skill's primary function is to generate documentation (Markdown files) based on the repository's file structure and contents. It does not exhibit any malicious patterns such as remote code execution, data exfiltration, or credential theft. All operations are limited to the local filesystem within the current repository context.
[SAFE]: The skill follows a principle of least privilege, explicitly forbidding the writing of application code or the execution of destructive commands. It is restricted to reading the project and writing documentation to the root or a dedicated docs/ directory.
[SAFE]: Transparency and user oversight are incorporated into the workflow. The skill includes a discovery phase to detect existing documentation and an interview phase to gather specific business context from the user before generating any content, ensuring the agent does not overwrite existing work without explicit permission.

agent-context