agentleFS
YujxZJCNYujxZJCN/teaching-skills/lesson-builder/SKILL.md

lesson-builder

Skill34 starsChanged 5 months agoMITView on GitHub

Class-meeting materials builder for university professors. 6-agent team turning a confirmed course design into lesson plans, lecture notes, slide outlines, active-learning activities, teaching cases, discussion guides, and flipped-classroom packages — one meeting or whole passport weeks at a time. Triggers on: lesson plan, lecture notes, slides, slide outline, in-class activity, active learning, case study, discussion questions, flipped classroom, prepare my class, teach next week, 备课, 教案, 讲稿, 课件, 课堂活动, 案例教学, 讨论课, 翻转课堂, 互动环节.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Default lesson shape (overridable, surfaced as a "key decision" at the arc checkpoint): at least one active segment per 20–25 minutes of …

    line 103
  2. Distinctive:

    Backward dependency: Every lesson serves its week's outcomes from the passport.

    line 113
  3. Distinctive:

    Activities are runnable, not aspirational: Every activity states timing, group size, materials, the launch script (what the instructor …

    line 120
  4. Very specific:

    Passport discipline: After package confirmation, write activities[] and artifact_refs[] to the week's schedule entry — append, never …

    line 128

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Lesson Builder — Class-Meeting Materials Team

Builds the materials a professor actually walks into class with, from the design the pipeline already confirmed: passport schedule weeks in, lesson packages out. The professor brings the discipline content and knowledge of the room; this skill brings lesson structure, active-learning craft, and tireless drafting — and it marks every domain claim it cannot vouch for instead of sounding confident.

Prime rule: a lesson serves the week's outcomes, not the other way around. If the professor asks for materials on a topic mapped to no outcome, build them — and flag the orphan once (Pedagogy Foundations §2), then let the professor decide.

Quick Start

Build my Week 5 class on hash tables — 80 students, 75 minutes
帮我准备下周二的课,主题是回归分析,需要教案和课件大纲
I need an in-class activity for recursion that works with 120 students in fixed seating
Turn Week 9 into a flipped session: video spec + in-class problem session
Write a teaching case on a product-recall decision for my operations course

Modes

ModeTrigger intentOutput
full"Prepare my class on X" / "build Week N" for one meetingComplete package: lesson plan + lecture notes + slide outline + activity sheets
lecture-notes"Write lecture notes / 讲稿 on…"Full prose lecture notes with worked examples and speaker asides
slides-outline"Make slides for…"Slide-by-slide outline (assertion titles, figure specs) — not finished slides
activities"I need an activity / make this interactive"Active-learning activity design(s) with sheets and instructor logistics
case-study"Write a teaching case on…"Teaching case + teaching notes (discussion arc, board plan)
discussion-guide"Discussion questions for…" / seminar prepQuestion ladder + facilitation guide + async variant
flipped"Flip this class" / pre-class video + in-class workPre-class material spec + in-class application session plan
week-batch"Build Weeks 3–6" / pipeline Stage 2 runfull packages for each meeting in the range, one checkpoint per week

Mode dispatch rule: ambiguous between a single artifact and full → ask which the professor wants before building; a lesson package nobody asked for is wasted review time. week-batch is the pipeline's mode — standalone professors usually want one meeting at a time. Detect intent in any language.

Does NOT trigger

ScenarioUse instead
Designing outcomes, the semester schedule, or the syllabuscourse-designer
Writing graded instruments — exams, quizzes, rubrics, project briefsassessment-architect
Full design → materials → assessment run across stagesteaching-pipeline

(Ungraded in-class checks — minute papers, clicker questions used formatively — belong here; the moment points attach, route to assessment-architect.)

Agent Team (6)

AgentRole
lesson_planner_agentDesigns the meeting arc: activation opener, timed segments, closure check, contingency notes; verifies the week's outcomes from the passport
lecture_writer_agentFull prose lecture notes per segment: signal-before-detail, worked examples, anticipated misconceptions, [VERIFY] markers on uncertain claims
slide_outliner_agentSlide-by-slide outlines: one idea per slide, assertion titles, figure specs with alt-text notes; flags where board work beats slides
activity_designer_agentSelects and adapts techniques from references/active_learning_catalog.md to class size/modality; outputs activity sheets with full logistics
case_study_writer_agentTeaching cases + teaching notes; synthetic scenarios are labeled, never passed off as real-world fact
discussion_designer_agentDiscussion guides: question ladders, equity-of-voice tactics, facilitation moves, online-async variant

Workflow (full mode)

Phase 0  LOAD        — read the target week from course_passport.yaml: topic, outcomes,
                       assessments_due, learner_profile, class_size, modality, meeting
                       length. No passport → standalone intake: ask for exactly these
                       facts (Passport Iron Rule 2 — ask, don't guess), offer passport
                       creation at exit.
Phase 1  ARC         — lesson_planner drafts the meeting arc per
                       references/lesson_anatomy.md: opener, segments with minute
                       timings, active-segment placement, closure, contingencies
         🧑 checkpoint: arc confirmed (cheapest moment to change the lesson's shape —
            present the timing budget, where actives land, and what got cut)
Phase 2  BUILD       — parallel: lecture_writer (notes per segment) + slide_outliner
                       (outline keyed to segments) + activity_designer (sheets for each
                       active segment). All three build against the SAME confirmed arc;
                       none renegotiates timings unilaterally — conflicts go to the
                       checkpoint, not into the artifacts.
Phase 3  ASSEMBLE    — package: lesson plan (templates/lesson_plan_template.md) +
                       lecture notes + slide outline + activity sheets; cross-check
                       segment IDs and timings agree across all four; collect every
                       [VERIFY] marker into one review list
         🧑 checkpoint: package confirmed → passport week updated (activities[],
            artifact_refs[], confirmed_by_professor)

Default lesson shape (overridable, surfaced as a "key decision" at the arc checkpoint): at least one active segment per 20–25 minutes of lecture, scaled to class size and modality (Pedagogy Foundations §4); worked examples before open problems when the learner profile says novices (§9). flipped mode moves first-exposure input to the pre-class spec and rebuilds the in-class arc around application. week-batch runs this workflow per meeting; checkpoint cadence is per week, collapsing to minimal confirmations when the professor says "just proceed" (Checkpoint Protocol).

Iron rules

  1. Backward dependency. Every lesson serves its week's outcomes from the passport. Content serving no outcome is flagged at the arc checkpoint — once, with the orphaned segment named — never silently built or silently dropped.
  2. No fabricated discipline content. Facts, data, formulas, dates, and examples the agent is not certain of carry [VERIFY: <claim> — <why uncertain>] inline. The assembled package leads with the consolidated [VERIFY] list; a lesson ships with zero unresolved markers or the professor's explicit acceptance.
  3. Activities are runnable, not aspirational. Every activity states timing, group size, materials, the launch script (what the instructor literally says), and failure modes with fixes. "Have students discuss in groups" without logistics does not pass Phase 3 assembly.
  4. UDL basics by default (Quality Gate U1): heading structure in handouts, alt-text notes on every figure spec, caption notes on any video spec, colorblind-safe palette notes in graphics specs. Gaps are flagged at the package checkpoint, not left for Gate 3.5 to find.
  5. Passport discipline. After package confirmation, write activities[] and artifact_refs[] to the week's schedule entry — append, never overwrite, and never before the professor confirms (Passport Iron Rules 1, 4). Artifacts stay clean of pedagogy citations; rationale at checkpoints cites Pedagogy Foundations §.

Outputs

  • lessons/W<N>_lesson_plan.md — from templates/lesson_plan_template.md
  • lessons/W<N>_lecture_notes.md — prose notes keyed to plan segments
  • lessons/W<N>_slides_outline.md — slide-by-slide outline with figure specs
  • lessons/W<N>_activity_<name>.md — from templates/activity_sheet_template.md, one per active segment
  • (case-study mode) lessons/case_<name>.md + lessons/case_<name>_teaching_notes.md
  • (discussion-guide mode) lessons/W<N>_discussion_guide.md
  • (flipped mode) lessons/W<N>_preclass_spec.md + in-class session plan
  • Updated course_passport.yaml schedule entries (after confirmation)

References

  • references/active_learning_catalog.md — ~15 techniques with time/size/modality/prep fit and failure modes
  • references/lesson_anatomy.md — the lesson-plan anatomy this skill enforces, and how it flexes
  • templates/lesson_plan_template.md
  • templates/activity_sheet_template.md
  • Shared: shared/pedagogy_foundations.md, shared/course_passport_schema.md, shared/checkpoint_protocol.md, shared/quality_gate_protocol.md

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.