The Agent Skills Directory

[COMMAND_EXECUTION]: The entry-point script scripts/generate_schematic.py executes a secondary logic script via subprocess.run. It correctly passes arguments as a list rather than a single string, which prevents command injection by ensuring user input is not interpreted by a shell.
[EXTERNAL_DOWNLOADS]: The skill communicates with OpenRouter (openrouter.ai) to generate diagrams and perform quality reviews. These network operations are necessary for the skill's primary function and target a well-known service. It requires the standard requests library.
[CREDENTIALS_UNSAFE]: The skill correctly manages API keys by retrieving them from environment variables or .env files. The documentation provides clear instructions for secure setup and lacks any hardcoded secrets or unsafe credential handling patterns.
[PROMPT_INJECTION]: The skill ingests user descriptions for diagrams and incorporates them into prompts for AI models. This indirect data flow is used solely for image generation and does not provide an avenue for system compromise or safety guideline bypass.

scientific-schematics