The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Node.js spawn and Python subprocess.run to handle project tasks such as next build, npm run validate, and npx shadcn@latest. These are standard operations for a static site generator and authoring framework, used here to materialize the application and manage components locally.
[EXTERNAL_DOWNLOADS]: Sub-skills within the engine reference downloading design assets from well-known services such as Pexels and Unsplash. Additionally, the tool provides integration with the Google Gemini API for automated design briefs and logo generation. All package dependencies are standard libraries from official registries.
[CREDENTIALS_UNSAFE]: No hardcoded secrets, tokens, or private keys were found. API key usage is instructed via environment variables with clear placeholders like your-key provided for the user's own configuration.
[PROMPT_INJECTION]: The skill instructions for AI agents are well-structured and prioritize the preservation of human-authored source material. It explicitly prohibits agents from overwriting raw evidence and requires user authorization for external research, mitigating unauthorized behavior overrides.
[DATA_EXFILTRATION]: The engine operates as a local-first application. No unauthorized network operations or patterns involving the transmission of sensitive local files to remote servers were identified.

lecture-site-engine