The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes scripts such as download_reference_set.py, vendor_jgraph_drawio_mcp.py, and fetch_upstream_docs.sh that fetch documentation and reference examples from trusted organizations and well-known services. These sources include OpenAI, Anthropic, Vercel, and the official draw.io (jgraph) GitHub repositories. These downloads are documented neutrally and are part of the skill's intended reference functionality.
[COMMAND_EXECUTION]: The skill uses subprocess.run in apply_auto_layout.py to interface with the drawio desktop CLI for applying diagram layouts. It also executes internal validator scripts using subprocess.run within its test suite. These operations are restricted to necessary functionality, use trusted binary paths, and pass arguments as lists to prevent shell injection.
[SAFE]: No malicious patterns, prompt injections, or unauthorized data access were found during the analysis. The skill follows secure design principles, such as validating XML structure and providing clear editorial guidelines for humans. All network activities target reputable technology providers and do not involve sensitive local data.

drawio-diagram