Workshop Style logo

Workshop Style

OrganizationPopular
THU-MAIC
workshop-style

Hands-on workshop — the learner works through the course instead of watching it, with an exercise or a check every couple of pages, short slides that only set up the next task, guiding narration that asks more than it tells, and a closing page where the pieces come together in one small project. Use when the ask is a workshop, a bootcamp session, 「动手」「实操」「练中学」「边做边学」, or a skill the learner is meant to leave holding; also use when the user names this style. Not for a systematic explanatory course (that is `lecture-style`) and not for occupational procedure training with tools and safety gates (that is `vocational`).

Overview

PublisherTHU-MAIC
RepositoryOpenMAIC
Skill nameworkshop-style
Stars
37.6K
Forks
5.9K
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by THU-MAIC on GitHub. Read the source before you install it.

Installation

Install the Workshop Style AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/THU-MAIC/OpenMAIC.git /tmp/OpenMAIC
mkdir -p .claude/skills
cp -r /tmp/OpenMAIC/skills/agent-runtime/workshop-style .claude/skills/workshop-style
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Workshop Style in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Workshop Style on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Workshop Style is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Workshop design

You are designing a workshop. Nothing is explained for longer than it takes to set up the next thing the learner does, and every idea is followed immediately by a page where they try it. The learner leaves having built something, not having been told about it.

stage-design still governs how the stage is built — outline, roster, one page per generate_scene, audio before done. This skill governs what the stage contains and how it sounds.

The shape

  • Scene 1 is a slide, and it is short: what the learner will have made by the end, and the first thing they are about to try. No agenda, no history of the field.
  • Every concept is followed at once by a hands-on page. The rhythm is concept → do it, concept → do it: roughly one interactive or quiz page for every two pages in the course. If two explanatory pages ever sit next to each other, one of them is unnecessary.
  • interactive pages are the exercises. Each isolates one thing to build, vary or trace. Consecutive interactive pages must not reuse the same widgetType — as the learner moves from trying to predicting to applying, the modality moves with them.
  • quiz pages are quick self-checks, not exams — placed right after a hands-on page to make the learner state what they just observed. Short, two or three questions.
  • The last page is a small integrating project: one interactive page whose task needs several of the workshop's pieces at once, with the earlier exercises as its parts. Not a summary slide. This is the page the learner screenshots afterwards.

Slides are task briefs

A slide in a workshop exists to make the next exercise possible. Keep it to two or three keyPoints — the rule, the shape, the one gotcha — and put everything else in the exercise. A workshop slide that could be read on its own as a lesson is too long: cut it and let the interactive page teach.

widgetOutline

Every interactive page must carry a populated widgetOutline, always including concept — the one thing this exercise is about. An interactive page with an empty widget outline degrades into a generic page, which in this style takes the course's whole point with it.

  • simulation: concept, keyVariables (the actual things the learner moves)
  • code: concept, language
  • diagram: concept, diagramType, nodes
  • game: concept, gameType, challenge, playerControls
  • visualization3d: concept, visualizationType, objects, interactions

keyPoints on an interactive page describe the task and what doing it reveals — what the learner changes, what they should notice, what conclusion that forces. Not facts to be read aloud.

Narration is facilitation

This is where the style is heard, so it is the part to get right.

  • Lines are short — one or two sentences. Nobody in a workshop listens to a paragraph while a task is waiting on screen.
  • Ask more than you tell. Most narration lines on a hands-on page are questions or instructions: 「先把这个值调到最大,看看会发生什么」「你觉得下一步 会往哪边偏?」「试完再往下看。」 Statements are reserved for the moment after the learner has already seen the effect.
  • Hand the work over and get out of the way. Say what to do, then stop. A narration that explains the answer before the exercise runs has cancelled the exercise.
  • Two voices, in dialogue. The teacher sets the task; the assistant co-facilitator reacts, tries the wrong thing on the learner's behalf, or asks what everybody is wondering. Alternating short turns, not monologue.
  • Encouragement is fine and belongs after an attempt, tied to something specific — 「注意你刚才那一下,曲线立刻翻过去了」 rather than 「太棒了」.

The roster

The runtime allows exactly one teacher, so the dialogic pair is a lead teacher plus a co-facilitator assistant — not two teachers. Write the roster with set_roster: a hands-on practitioner who talks in tasks, an assistant who is warm, quick and unafraid to be wrong out loud, and keep the roster small enough that turns stay readable. Put 「short turns, question-first, encouraging」 in the personas and a lively pace in voiceDesign.delivery.

Carrying the style into the generators

There is no outline generator: you plan the course in the conversation, then call create_stage and one generate_scene per page with an explicit brief. Page content and narration are written by separate calls that see only the page's brief, the page content and the roster. So the style has to be carried there explicitly:

  • put the concept → exercise cadence and the closing project into each page's generate_scene brief;
  • put the two-voice facilitation manner into each persona of set_roster;
  • pass the narration directive for each page in generate_scene.materialFacts — e.g. 「旁白为引导式短句,一到两句一条,提问多于陈述」「本页先让学员动手, 结论留到操作之后」. Naming this skill in a prompt does nothing; only the text you place in those fields is seen.

Scene naming

Titles name what the learner does, in the imperative.

  • Good: 「把这段循环改成能跑的」「调参数,找出它崩掉的那一刻」「用今天的三块拼一个 能用的小工具」
  • Bad: 「循环的概念」「参数敏感性分析」「综合练习」

If the request is not a workshop

If the subject has nothing to practise — a historical narrative, an appreciation course, a body of knowledge the learner only needs to understand — say so in one sentence and use lecture-style instead. Bolting exercises onto a topic with no task produces busywork, which learners recognize faster than anything else.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Workshop Style AI skill do?

Hands-on workshop — the learner works through the course instead of watching it, with an exercise or a check every couple of pages, short slides that only set up the next task, guiding narration that asks more than it tells, and a closing page where the pieces come together in one small project. Use when the ask is a workshop, a bootcamp session, 「动手」「实操」「练中学」「边做边学」, or a skill the learner is meant to leave holding; also use when the user names this style. Not for a systematic explanatory course (that is `lecture-style`) and not for occupational procedure training with tools and safety gate...

Why use Workshop Style on TypingMind?

Because you install it once and use it with any model. Workshop Style is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Workshop Style in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/THU-MAIC/OpenMAIC/tree/main/skills/agent-runtime/workshop-style. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Workshop Style?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Workshop Style?

As many as you like. As long as a model supports skills, you can use Workshop Style with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Workshop Style AI skill free?

Yes. It is published on GitHub by THU-MAIC under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇