The Agent Skills Directory

[SAFE]: The skill provides guidelines for the agent to generate visual responses such as call trees, component trees, and file trees instead of verbose text.
[COMMAND_EXECUTION]: The skill includes instructions to use the open command on HTML files generated in the system's temporary directory ($TMPDIR) to display complex visualizations to the user. This is an intended functional capability and does not involve network exfiltration or sensitive file access.
[INDIRECT_PROMPT_INJECTION]: The skill processes user queries and project file structures to generate its views.
Ingestion points: Project file names, paths, and source code content from the current working directory.
Boundary markers: None explicitly defined in the prompt instructions.
Capability inventory: File system read access, file write access to $TMPDIR, and command execution (open).
Sanitization: No explicit sanitization of file paths or names before rendering in diagrams. The risk is assessed as safe as the output is intended for user visualization and the tool does not automatically chain into other high-privilege actions based on the generated text.

show-me