The Agent Skills Directory

[COMMAND_EXECUTION]: The utility script render-graphs.js uses child_process.execSync to run the system command dot for rendering diagrams. The script passes diagram content to the command via stdin, which is a secure implementation that avoids shell injection risks.
[PROMPT_INJECTION]: The skill methodology teaches the use of "Authority" and "Commitment" principles (such as using "YOU MUST" or "No exceptions") to ensure agents follow critical rules. These are instructional techniques intended to increase process reliability and are not used as malicious safety bypasses.
[INDIRECT_PROMPT_INJECTION]: The render-graphs.js script provides a surface for processing external documentation that could contain untrusted instructions.
Ingestion points: Reads SKILL.md files from a user-provided directory path.
Boundary markers: None.
Capability inventory: File system access and execution of the local dot binary.
Sanitization: Content is extracted from markdown code blocks and passed directly to the renderer without additional validation.

writing-skills