build-course-fill-in-workbook
Skill0 starsChanged 2 months agoMITView on GitHub
Turn course notes, slides, or homework into a printable practice workbook with examples first, less help over time, and answers kept behind a clear STOP page. Use when the user wants safe practice rather than answers to live graded work.
In this repository7 other files it gives its agents.
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Use available context and safe, reversible defaults before asking for input.
- Distinctive: line 36
Keep a simple record of where each claim came from: seen in a source, calculated, a teaching choice, or uncertain.
- Distinctive: line 44
Follow the rubric's progression: worked example, then practice with fading help, then one independent task the learner can finish with no …
- Distinctive: line 47
Keep learner pages before a prominent STOP page and every answer after it, or in a separate answer file.
- Distinctive: line 58
Check every diagram and step-by-step trace against the source or the running code.
- Distinctive: line 65
Follow the layout-and-QA reference: render every page, inspect each one at a readable size, confirm the STOP/answer boundary, and rerun the …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Build Course Fill-In Workbook
Turn the user's course material into a printable practice workbook: worked examples first, then practice with less and less help, then one independent task, with every answer kept behind a clear STOP page.
Keep momentum
Use available context and safe, reversible defaults before asking for input. Bundle non-blocking questions and continue useful work. Interrupt only when a missing choice would change the result, privacy, cost, authority, safety, or something hard to undo. Never repeat an answered question.
Read the right reference first
- learning-design-rubric.md before planning the practice sequence.
- assessment-safety.md whenever the sources include quizzes, homework, tests, or unclear assessment status.
- layout-and-qa.md before laying out or checking the PDF.
- evidence-basis.md only when explaining why the learning design works.
- established-examples.md only when matching the shared workbook pattern or project layout.
Build from the user's material
- Collect the supplied notes, slides, transcripts, instructions, starter code, and visible tests. Keep every original file unchanged.
- Keep a simple record of where each claim came from: seen in a source, calculated, a teaching choice, or uncertain. If the course already has a shared source folder (approved sources, key terms and pronunciations, and what is graded), reuse it so the workbook and any matching study podcast agree. Ask instead of guessing when a source is ambiguous.
- Name what the learner should be able to do afterward — explain, trace, calculate, debug, or write — and open with one small concrete example, defining each new term where it first appears.
- Follow the rubric's progression: worked example, then practice with fading help, then one independent task the learner can finish with no answers in sight. Skip support the learner has already shown they do not need.
- Keep learner pages before a prominent STOP page and every answer after it, or in a separate answer file.
- End with a mixed checkpoint, and list each topic with a suggested date to review it again in a later session. A same-session checkpoint is not real spaced review, so do not describe it as one.
Check the content
- For code or calculations, build the complete solution separately, run the supplied tests plus sensible edge cases, and confirm the blanked version can be rebuilt from the answer key and still passes the same tests.
- Check every diagram and step-by-step trace against the source or the running code. Keep BEFORE and AFTER states clearly separate.
- Scan learner pages for leaked answers, hidden text, and live-assessment wording.
Check the finished PDF
Follow the layout-and-QA reference: render every page, inspect each one at a readable size, confirm the STOP/answer boundary, and rerun the code checks after the final edit. If the user asked for a copy somewhere else, confirm the delivered copy matches the checked one.
Report what was checked and anything that was not. "Generated" is not "verified," and never invent measurements of how the learner did.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

