The Agent Skills Directory

[SAFE]: The skill's primary function is documentation search. It operates by reading local Markdown files in the references/ directory and does not include any executable scripts or binary code.
[SAFE]: The workflow defined in SKILL.md uses a normalization process for user input that prevents directory traversal and restricts file access to the specific documentation subdirectory.
[SAFE]: While the documentation content includes technical examples such as curl commands and configuration snippets, these are provided as plain-text documentation for the user and are not automatically executed by the agent.
[SAFE]: The skill uses clear instructions to ground its responses in the provided documentation files and includes placeholders for user API keys in examples, adhering to standard documentation security practices.

cursor-docs