The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes scripts like scripts/fetch_images.py to download image assets and logos from well-known sources including Wikimedia Commons, Unsplash, Pexels, and official brand websites. It also references scripts/tts-doubao.mjs for interacting with the Doubao TTS API.
[COMMAND_EXECUTION]: Extensive reliance on local shell and script execution is present for tasks such as media processing (ffmpeg, ffprobe), PDF/PPTX export (scripts/export_deck_pdf.mjs, scripts/export_deck_pptx.mjs), and automated screenshots or video rendering using Playwright.
[DYNAMIC_EXECUTION]: The core functionality involves dynamically generating HTML/React code and executing it within a browser environment (rendered via Playwright) to produce visual mockups and animations. This allows for runtime execution of generated code to facilitate video export and previewing.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect injection as it is designed to ingest and process data from external sources such as WebSearch results, user-provided codebases, and brand assets. These external inputs are incorporated into the prompt and code generation pipeline. The risk is addressed internally by strict instructions for fact-checking and asset verification.

huashu-design