LessonLab is a collection of prompts, catalogs, and templates for teachers who want to plan better lessons with AI. It works with ChatGPT, Claude, Copilot (in ChatGPT mode), and Gemini. You do not need to be a developer, and you do not need a GitHub account to use it.
If you have a GitHub account, fork this repository to your own account. That way you can pull updates as the prompts improve.
- Go to
https://github.com/goneil78-coder/LessonLab - Click Fork (top right)
- Open any
.mdfile, read it, and copy the prompt inside
You do not need to clone anything to your computer. Every file is readable directly in your browser on GitHub.
If you accessed this through a SharePoint link or received it as a file, you already have what you need.
- Unzip the file if it is still zipped
- Open the folder
- Open
README.md(this file) in any text editor, Notepad, or Word - Open any
.mdfile inprompts/and copy the prompt inside
That is it. You copy prompts out of these files and paste them into ChatGPT, Claude, Copilot, or Gemini. Nothing to install.
LessonLab is not just a pile of prompts. Used well, it works as a structured lesson-planning skill that your AI runs. To get that experience, orient the AI to the folder before you ask it to do anything specific.
There is a file at the root called SYSTEM.md written directly to the AI. It explains what LessonLab contains, how the pieces fit together, and how to adapt to whichever platform you are on (Claude Desktop, ChatGPT Projects, Custom GPT, Claude.ai web, Copilot, Gemini).
Easiest path. Open prompts/first-prompt.md. It is a short copy-paste block that tells your AI to read SYSTEM.md and ask you the right starting question. Paste it once on day one. From then on, your AI knows what LessonLab is and routes you to the right prompt based on your project.
For Claude Desktop or ChatGPT with Projects. Create a new Project, upload the LessonLab files, and paste the contents of SYSTEM.md into the Project's Instructions field. Every chat in that Project starts pre-oriented. You do not need to use first-prompt.md again once the Project is set up.
For ChatGPT Custom GPTs. You can publish "LessonLab Assistant" as a Custom GPT that colleagues install with one click. Instructions field = SYSTEM.md contents. Knowledge files = the prompt and catalog files in this repo. Share the GPT link with your department.
Full platform-by-platform setup (including Claude.ai web without Projects, Copilot in ChatGPT mode, and Gemini for Workspace) is documented inside SYSTEM.md itself.
This sequence takes you from a topic to a lesson plan ready to teach. Steps 1 and 2 are optional but strongly recommended for a better result.
Step 1: Research the topic (5 minutes)
Open prompts/research-guide.md. Follow the five-step sequence inside. It walks you through asking your AI for a background overview, key misconceptions, counterarguments, and curriculum connections. By the end you will have a compact research brief to paste into the next step.
Step 2: Choose your lesson plan prompt (2 minutes)
Pick the right prompt for your subject:
- General subjects:
prompts/lesson-plan-general.md - Theory of Knowledge:
prompts/lesson-plan-tok.md - IB English B / Language B:
prompts/lesson-plan-english-b.md
Open the file and read the first section. It tells you what to prepare before you paste the prompt.
Step 3: Paste the prompt and your research brief (10 minutes)
Open a new chat in your AI tool. Paste the lesson plan prompt. Then paste your research brief from Step 1. Send it. Read the output. Ask follow-up questions to push the plan further.
Step 4: Run the quality check (3 minutes)
Open prompts/devils-advocate.md. Paste your draft lesson plan where indicated and send it. The AI challenges each recommendation across five dimensions: context fit, implementation cost, pedagogical trade-offs, evidence strength, and preservation of your original intent. Use the output to decide what to keep, modify, or drop.
That is a complete planning cycle. Optional steps 5 and 6 below extend it.
Step 5 (optional): Check for AI tells in written materials
If your lesson plan produces text you will share with students or publish, open prompts/ai-tells-scan.md. If you do not yet have a personal tell list, run prompts/build-your-ai-tells.md first. It takes about 15 minutes and you only do it once.
Step 6 (optional): Validate multiple artifacts together
If you have generated a lesson plan and a slide deck (or any two related outputs), open prompts/validate-outputs.md. It checks that the two artifacts are internally consistent with each other.
These prompts are organized by the stage in a planning workflow where you would use them.
| File | What it does |
|---|---|
prompts/first-prompt.md |
Day-one paste-in. Tells your AI to read SYSTEM.md and orient itself before doing anything else. Use once per new chat or new platform. |
prompts/setup-interview.md |
Interviews you about a recurring project and produces a filled-in SESSION.md dashboard file. Use this when starting a project you will return to over weeks or months. |
prompts/post-session-update.md |
Updates an existing SESSION.md at the end of a working session. Keeps the dashboard current so the AI has accurate context next time. |
| File | What it does |
|---|---|
prompts/research-guide.md |
Five-step structured research sequence. Takes you from a blank topic to a usable brief covering key ideas, misconceptions, counterarguments, and curriculum connections. |
| File | What it does |
|---|---|
prompts/lesson-request-template.md |
Ten-field input template you fill in before any lesson plan prompt. Stops "a lesson on [topic]" from producing generic output. Pair this with one of the three plan generators below. |
prompts/lesson-plan-general.md |
Full lesson plan generator for any subject. Produces a structured plan including learning objectives, activities, differentiation, and assessment. |
prompts/lesson-plan-tok.md |
Lesson plan generator for IB Theory of Knowledge. Calibrated to TOK knowledge questions, the exhibition and essay requirements, and core / optional theme structure. |
prompts/lesson-plan-english-b.md |
Lesson plan generator for IB English B and Language B. Calibrated to acquisition pedagogy, the four skills, text types, and assessment objectives. |
Options first, full plan second. Before asking for a full plan, paste the filled request template and ask "give me two to four possible directions this lesson could take, one sentence each." Pick the strongest. Only then ask for the full plan on that direction. Costs 60 seconds, beats marrying a mediocre first draft.
| File | What it does |
|---|---|
prompts/devils-advocate.md |
Adversarial review of a draft lesson plan. Challenges each recommendation across five dimensions and returns verdict icons (strong / conditional / weak) for prioritization. |
prompts/validate-outputs.md |
Cross-artifact alignment check. Confirms that two related outputs (lesson plan and slides, for example) are consistent with each other before you teach. |
| File | What it does |
|---|---|
prompts/build-your-ai-tells.md |
Interview-based process for discovering your personal AI tells. Compares your original writing to AI-generated versions of the same text to surface the patterns specific to you. Run once; save the output. |
prompts/ai-tells-scan.md |
Scans a draft against your personal tell list and returns a tier-ordered findings report with specific edit suggestions. Requires your tell list from build-your-ai-tells.md. |
The catalogs are reference files. You feed them to your AI alongside a lesson plan prompt to give the AI richer options to draw from.
| File | What it contains |
|---|---|
catalogs/harvard-project-zero-routines.md |
38 thinking routines from Harvard Project Zero (See-Think-Wonder, Claim-Support-Question, Headlines, and more), each with a brief description and suggested use. |
catalogs/pedagogical-frameworks.md |
Synthesis of five evidence-based frameworks: Rosenshine's Principles, Willingham's cognitive science of learning, MARGE (memory consolidation), Dylan Wiliam on formative assessment, and Ron Berger on critique culture. |
catalogs/questioning-feedback.md |
20 questioning and feedback techniques for classroom use, covering cold calling, wait time, probing sequences, written feedback formats, and exit ticket design. |
catalogs/english-b-acquisition.md |
Course design reference for IB English B and Language B, covering acquisition theory, skill integration, text types, and assessment objectives. |
catalogs/explaining-modeling.md |
10 explaining and modelling techniques including worked examples, think-alouds, live modelling, and error analysis. |
catalogs/other-lesson-routines.md |
18 additional lesson routines covering lesson openers, pair and group structures, plenaries, and reflection activities. |
catalogs/mode-b-teaching.md |
8 Mode B teaching techniques plus the ADAPT process for moving between structured (Mode A) and more open-ended (Mode B) pedagogy. |
How to use catalogs with a prompt: Before sending a lesson plan prompt, paste the relevant catalog content into your chat and say something like "I am also attaching a catalog of thinking routines. Draw from these when selecting activities." The AI will then make choices from the actual options rather than its generic defaults.
| File | What it contains |
|---|---|
examples/cohort-session-redesign-example.md |
A real filled-in SESSION.md from a teacher's multi-session work redesigning one session of a cohort training programme. Attendee names have been anonymised to role labels. Shows what a dashboard looks like after it has been running through several working sessions, with Known Issues and Session Log genuinely accumulated. |
| File | What it contains |
|---|---|
style/folding-thoughts-reference.md |
An example style reference file showing typography, colour palette, tone, and formatting conventions for one teacher's lesson materials. This is a template to replace, not a file to use as-is. |
How to create your own style reference: Copy style/folding-thoughts-reference.md to a new file (for example style/my-school-style.md). Replace the contents with your school's house style, your department conventions, or the register you use in your own materials. Feed this file to your AI whenever you are generating slides, handouts, or written outputs that need to look consistent.
Start with prompts/lesson-plan-tok.md rather than the general prompt. It understands TOK knowledge questions, the difference between exhibition and essay preparation, and the structure of the core and optional themes. Pair it with catalogs/harvard-project-zero-routines.md for thinking routines that align naturally with TOK inquiry.
Start with prompts/lesson-plan-english-b.md. It is calibrated to acquisition pedagogy rather than literary analysis, covers all four skills, and accounts for the IB English B text type requirements. Pair it with catalogs/english-b-acquisition.md for course design reference.
Use prompts/lesson-plan-general.md as your base. The catalogs in catalogs/ apply across subjects. Before running the plan prompt, decide which catalog is most relevant for your particular lesson goal (a questioning-heavy discussion lesson might draw from catalogs/questioning-feedback.md; a modelling-heavy skills lesson might draw from catalogs/explaining-modeling.md) and paste it into the chat.
All prompts in LessonLab are written to work with any of the four major chat AI tools. There are a few practical differences worth knowing.
ChatGPT (Projects feature): If you use ChatGPT Projects, upload your SESSION.md and any relevant catalogs directly into the project. They persist across sessions without you needing to paste them each time. This is the easiest setup for recurring workflows.
Claude (Projects feature): Claude Projects works the same way. Upload SESSION.md and your style reference into the project context. Claude tends to handle long structured documents well, so you can include multiple catalogs at once without a significant drop in quality.
Copilot in ChatGPT mode (Microsoft 365): Copilot in ChatGPT mode does not have a persistent projects feature in the same way. Paste your SESSION.md and the relevant catalog at the top of each chat. If your organisation uses SharePoint, you may be able to reference files directly via the file picker, depending on your tenant configuration.
Gemini (Google Workspace): Gemini for Google Workspace can read Google Docs. Consider keeping your SESSION.md and key catalogs as Google Docs in a shared Drive folder. You can then instruct Gemini to read them directly rather than pasting the text.
General rule: Longer context (more text pasted in) generally improves output quality across all four tools. If a prompt is producing generic results, try adding the relevant catalog to your chat before sending the lesson plan prompt.
- Fork this repository to your own account
- Make your changes (new prompt, improved catalog, corrected example)
- Open a Pull Request against the main repository with a brief description of what you changed and why
Small, focused contributions are easiest to review. One new prompt per PR, or one catalog update, rather than several changes at once.
Open an issue at https://github.com/goneil78-coder/LessonLab/issues (GitHub issues are readable by anyone without an account; you only need an account to post). If you cannot use GitHub at all, leave a comment on the SharePoint document where you received this, or ask a colleague with a GitHub account to open an issue on your behalf. Include the file name you are suggesting changes to, the specific text you would change, and what you would replace it with. That level of detail makes it straightforward to apply.
What good contributions look like: A prompt that has been tested in at least one real lesson. A catalog addition that cites where the technique comes from. A correction to a factual error in a catalog. A new SESSION.md example for a subject not currently represented.
LessonLab is published under the Creative Commons Attribution-ShareAlike 4.0 International license (CC BY-SA 4.0).
You are free to share (copy and redistribute in any medium or format) and adapt (remix, transform, and build upon the material) for any purpose, including commercial, as long as you:
- Attribute: credit this repository and link to the license
- ShareAlike: distribute any adapted material under the same license
Full legal text: https://creativecommons.org/licenses/by-sa/4.0/legalcode
Human-readable summary: https://creativecommons.org/licenses/by-sa/4.0/
This license covers all files in this repository unless a specific file states otherwise.