The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/render-drawio.mjs dynamically locates and loads the playwright or playwright-core library by querying global package manager roots (npm or pnpm) at runtime.\n- [COMMAND_EXECUTION]: The skill executes local Node.js scripts for diagram validation and rendering, and invokes shell commands (npm root -g, pnpm root -g) to resolve dependency paths.\n- [EXTERNAL_DOWNLOADS]: The skill fetches the official Draw.io static viewer script from viewer.diagrams.net during the rendering process. It also provides an option to fetch extended documentation from the official jgraph/drawio-mcp repository on GitHub.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data to generate XML diagrams which are subsequently handled by execution-capable scripts.\n
Ingestion points: Diagram content requested in user prompts is interpolated into generated mxGraphModel XML files.\n
Boundary markers: The skill uses scripts/validate-drawio.mjs to verify the structural integrity of the generated XML.\n
Capability inventory: Writing .drawio files, executing local node scripts, and performing network requests through a headless browser.\n
Sanitization: The validator script ensures XML and mxGraphModel schema requirements are met, but does not filter semantic content within diagram labels.

drawio