The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/generate_schematic.py and scripts/validate_format.py scripts utilize subprocess.run to execute local Python scripts and common system utilities including pdfinfo and pdffonts. These operations are performed on local files to provide the documented functionality of image generation and document compliance checking.
[EXTERNAL_DOWNLOADS]: The scripts/generate_schematic_ai.py script communicates with openrouter.ai, a well-known service for accessing various AI models. It uses this connection to send diagram descriptions and receive generated images, which is the primary feature of the scientific schematic capability. It also uses Google's Gemini models (a trusted organization) for quality review.
[DATA_EXFILTRATION]: The skill transmits user-provided natural language descriptions of scientific diagrams to the OpenRouter API to facilitate image generation. This is an expected and documented behavior for an AI-integrated tool requiring external processing.
[PROMPT_INJECTION]: The scripts/generate_schematic_ai.py script incorporates user input into a prompt sent to an external LLM. It employs basic boundary markers (e.g., 'USER REQUEST:') to maintain instruction integrity, representing a standard and low-risk implementation for sub-agent interactions.

venue-templates