The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the subprocess module in its orchestration scripts (precheck.py, postcheck.py) to run validation tools, manage the TTS generation, and invoke the video rendering CLI (hyperframes). This is a necessary and core part of its automated video production pipeline.
[DYNAMIC_EXECUTION]: Several validation scripts (check_curves_rendered.py, check_render_overlap.py, etc.) dynamically generate Node.js scripts into temporary files and execute them. These drivers utilize puppeteer-core to load generated scenes in a headless browser and perform visual inspections of the rendered geometry and layout.
[EXTERNAL_DOWNLOADS]: The pipeline includes logic to download the KaTeX library from its official GitHub releases and the GSAP library from the NPM registry if they are not already present in the local assets. Additionally, the TTS script communicates with the well-known Aliyun DashScope service. These targets are all trusted, well-established technology providers.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (math problems provided via text or images). While this constitutes an attack surface, the risk is mitigated by a rigid, multi-step pipeline where intermediate outputs (problem text, solution analysis) are formatted into structured markdown and verified by multiple automated gates before any code is executed or rendered.

qwen-mm-plugins-edu-agent