The Agent Skills Directory

[SAFE]: The skill interacts with the local filesystem to read and write a configuration file at ~/.config/skills/config.json. This is used solely to store the directory path of a skills repository, which is a legitimate and common pattern for developer tools.
[SAFE]: The skill processes untrusted data by reading README.md and CONTRIBUTING.md from a user-specified repository.
Ingestion points: Reads files from the user-provided skillsRepo path.
Boundary markers: None explicitly defined in the prompt interpolation, but the instructions focus on reading conventions rather than executing commands from these files.
Capability inventory: The skill only performs file reads and writes related to its own configuration and the repository's documentation.
Sanitization: No specific sanitization is mentioned, but the instructions limit the scope of interaction to reading conventions.

writing-skills-readme-first