The Agent Skills Directory

[SAFE]: The skill includes a custom Node.js script scripts/diagram-check.mjs for verifying SVG diagrams. The script is self-contained, utilizes only standard Node.js modules, and performs purely structural and geometric analysis of XML and CSS content.
[COMMAND_EXECUTION]: The instructions require the agent to execute a local linting script to ensure diagrams meet the defined quality budget. This execution is scoped to artifacts generated by the agent and is essential to the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill processes user prompts and reads external design contracts to inform diagram generation, which constitutes a potential injection surface.
Ingestion points: User prompts, design contract files, and existing SVG/tldraw artifacts.
Boundary markers: The skill relies on a rigid editorial workflow and technical verification rather than explicit prompt delimiters.
Capability inventory: Shell execution for structural verification, file system operations for artifact management, and integration with the tldraw API.
Sanitization: The verification gate (diagram-check.mjs) validates the technical constraints and XML structure of output files, preventing the accidental execution of instructions embedded in diagram metadata.

khe-diagram