qwen-mm-plugins-edu-agent

Generate step-by-step Mandarin math tutorial videos from problems or images.

2.9k|185|Updated Jul 29, 2026
One-click install
npx skills add https://github.com/QwenLM/Qwen-MM-Plugins --skill qwen-mm-plugins-edu-agent-qwenlm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: qwen-mm-plugins-edu-agent
Source: https://github.com/QwenLM/Qwen-MM-Plugins/tree/main/src/capabilities/edu-agent/skill
Command: npx skills add https://github.com/QwenLM/Qwen-MM-Plugins --skill qwen-mm-plugins-edu-agent-qwenlm

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires dashscope, soundfile, numpy, requests, hyperframes, ffmpeg, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Creating narrated, animated math tutorial videos normally requires scripting, voiceover recording, animation work, and video editing. This Skill automates the entire pipeline: it takes a math or science problem (text, LaTeX, or images in an image_assets/ folder) and produces a finished MP4 tutorial video with Mandarin narration, animated SVG diagrams, KaTeX formulas, and synchronized captions. ## Core Features & Use Cases - End-to-end 7-step pipeline: problem extraction from images, solution analysis, Chinese teaching script writing, Qwen-TTS voice generation with per-sentence timestamps, storyboard design, HTML/GSAP component building, and deterministic rendering to MP4 via hyperframes. - 83 pre-built K12 visual components: cars, circuits, lenses, beakers, pulleys, graph axes, and more, plus an Aurora Scholar design system with enforced visual quality gates (precheck/postcheck scripts that validate fonts, captions, geometry, and rendered pixels). - Use Case: A teacher drops photos of a geometry proof problem into image_assets/ and asks for an explanation video; the Skill extracts the problem via vision, solves it, narrates it in Mandarin with qwen3-tts-flash, animates the proof step by step, and outputs a rendered MP4. ## Quick Start Ask the agent to create a step-by-step Chinese tutorial video for your math problem, either by pasting the problem text or placing problem images in an image_assets/ folder.

Frequently Asked Questions about qwen-mm-plugins-edu-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a math tutorial video from a problem image?

Place the problem images in an image_assets/ folder and invoke the skill. It reads the images with multimodal vision, extracts the problem into LaTeX, solves it, writes a Chinese narration script, synthesizes voice with Qwen-TTS, and renders an MP4 via hyperframes.

What tools are needed to render educational videos with this skill?

You need Node.js with npx for the hyperframes CLI, headless Chromium for rendering, Python 3 with dashscope, soundfile, numpy and requests for TTS, ffmpeg for loudness normalization, and a DASHSCOPE_API_KEY for the qwen3-tts-flash voice model.

Does the skill support languages other than Chinese for narration?

No. The pipeline is designed for Mandarin output: the teaching script is written in Chinese at 3.5-4.0 characters per second, math symbols are converted to spoken Chinese, and captions come from the original Chinese script.

Why does my rendered video show blank scenes or missing curves?

Blank scenes usually come from missing self-hosted GSAP, KaTeX, or fonts in dist/, CDN URLs, or CSS-hidden content. Missing curves happen when a hand-rolled renderer strips GSAP. Run scripts/precheck.py before rendering and scripts/postcheck.py after to catch these.

Can I use custom visual components not in the asset catalog?

Yes, but they must meet the quality fallback template: at least three gradient layers, inset shadows, a ground shadow, a glow halo, no CSS keyframes, and a minimum 30px size. Compare against the candle component, which is the quality reference.