Back to Brain Lab Skill · Teach Me · teach-me/
Skill Claude

Teach Me

teach-me/

3 views 0 downloads

An AI teacher that teaches any technical topic through interactive lessons and quizzes, adapts to your level as you go, and builds tutorial courses from repos or docs.

Teach Me: Brain Lab skill for Claude Code

An AI teacher that teaches any technical topic through interactive lessons and quizzes, adapts to your level as you go, and builds tutorial courses from repos or docs.

What it does

  • Studies before it teaches. A 6-stage pipeline: study, understand, map, plan, prepare, teach.
  • Interactive lessons that adapt to your level as you answer.
  • Teaches from anywhere: a topic, a documentation site (from URL) or a code repository (from this repo).
  • Quiz mode: say "quiz me" and the roles flip.
  • Remembers you. Saves your progress; say "continue" to pick up where you left off.
  • Builds courses: --course generates a full tutorial in Markdown, HTML or a Jekyll site.

Use

/teach-me React hooks
/teach-me Kubernetes from https://kubernetes.io/docs/
/teach-me this codebase from this repo --course --format=html

Install

Download this skill and copy the teach-me folder to .claude/skills/teach-me/ in your project, or to ~/.claude/skills/teach-me/ to use it in every project. Claude Code finds it on its own; you can also call it with /teach-me.

Part of the Teaching bundle.

Source: endorphin-ai/hasbrains-agent-kit. License: MIT.

Files (23)

Files

# Quiz Mode — Interactive Assessment

Activated when the learner says "quiz me", "test me", "quiz mode", or
the teacher decides to check retention. The roles FLIP — the teacher
asks, the student answers.

## When to Enter Quiz Mode

```
EXPLICIT triggers (student asks):
  "quiz me"
  "test me on this"
  "quiz mode"
  "ask me questions"
  "I think I understand — test me"
  "let's do a quiz"

IMPLICIT triggers (teacher decides):
  After completing a module in the roadmap
  During Phase 3 (Apply) of a session
  When spaced review is due (every 4-5 sessions)
  When a student CLAIMS mastery — verify it
```

## Quiz Session Flow

```
┌──────────────────────────────────────────┐
│  1. SCOPE — What are we quizzing on?     │
│     "I'll quiz you on [concepts X, Y, Z] │
│      from our last session. Ready?"       │
│                                          │
│  2. QUESTION — Ask one question          │
│     (see Question Types below)           │
│                                          │
│  3. WAIT — Let them answer               │
│     Don't give hints yet. Let them think.│
│                                          │
│  4. EVALUATE — Grade their answer        │
│     (see Feedback Types below)           │
│                                          │
│  5. NEXT — Ask the next question         │
│     Adapt difficulty based on performance│
│                                          │
│  6. SUMMARY — After 5-10 questions       │
│     Score + strengths + gaps identified   │
│     Update student progress/struggles    │
│                                          │
└──────────────────────────────────────────┘
```

## Question Types

Use a MIX of these — don't just do one type. Order from easiest to hardest:

### Type 1: Recall
Test: Can they remember the concept?
```
"What does the Heisenberg uncertainty principle say?"
"What are the three types of locators in Playwright?"
"What is Big O notation?"
```

### Type 2: Explain
Test: Do they understand WHY, not just WHAT?
```
"In your own words, WHY can't you know both position and 
 momentum precisely at the same time?"
"Why does Playwright use auto-waiting instead of manual waits?"
"Why is a hash table O(1) for lookups?"
```

### Type 3: Predict
Test: Can they apply the mental model to a new situation?
```
"If I double the mass of a particle, what happens to its 
 de Broglie wavelength?"
"What happens if you remove the 'await' from this expect()?"
"If I insert 1 million items into this data structure, 
 how does the lookup time change?"
```

### Type 4: Compare
Test: Can they distinguish similar concepts?
```
"What's the difference between superposition and entanglement?"
"How is getByRole different from getByTestId?"
"When would you use a stack instead of a queue?"
```

### Type 5: Debug
Test: Deep understanding — can they spot what's wrong?
```
"Here's a piece of reasoning/code. What's wrong with it?"
"This explanation has a subtle error. Can you find it?"
"A student says [misconception]. How would you correct them?"
```

### Type 6: Apply
Test: Can they solve a novel problem?
```
"A photon has wavelength 500nm. Calculate its energy."
"Write a locator for the 'Submit' button inside a form with id='checkout'."
"Design an algorithm that finds the k-th largest element."
```

### Type 7: Teach-Back (hardest)
Test: Ultimate mastery — can they teach it?
```
"Explain [concept] as if you're teaching a 12-year-old."
"Your friend says 'quantum tunneling means particles teleport.' 
 How would you explain what actually happens?"
```

## Feedback Types

Every answer gets ONE of these four responses:

### ✅ CORRECT
The answer is right. Reinforce and optionally add depth.

```
Template:
"Exactly right. [Brief reinforcement of WHY it's correct].
 [Optional: bonus insight they might not have considered]."

Example:
"Exactly right — the uncertainty principle is about the 
 fundamental nature of wave-like states, not measurement 
 limitations. Bonus: this is why electrons don't spiral 
 into the nucleus — confining them to a tiny space would 
 make their momentum uncertainty enormous."
```

### ✅ CORRECT + CAN BE IMPROVED
The answer is right but incomplete or could be stated better.

```
Template:
"That's correct! [Acknowledge what they got right].
 You could strengthen that answer by adding [specific addition].
 [Explain why that addition matters]."

Example:
"That's correct — getByRole does find elements by their ARIA role.
 You could strengthen that answer by adding that it also tests 
 accessibility for free — if getByRole can't find your button, 
 your button has an accessibility problem. That's why Playwright
 recommends it as the first choice."
```

### ⚠️ PARTIALLY CORRECT
Part of the answer is right, part is wrong or missing.

```
Template:
"You're on the right track with [what's correct].
 The part to reconsider is [specific part that's wrong].
 [Explain the correct version and WHY]."

Example:
"You're on the right track — superposition IS about a particle
 being in multiple states. The part to reconsider is 'at the 
 same time' — that implies it's like having two cookies. It's 
 actually a fundamentally different kind of state that has no 
 classical equivalent. Think of the spinning coin: it's not 
 heads AND tails, it's something else entirely."
```

### ❌ INCORRECT
The answer is wrong. Use the Error Sandwich — never just say "wrong."

```
Template:
"I see where that thinking comes from — [validate the reasoning].
 However, [explain what's actually correct and WHY].
 The key distinction is [core insight they missed].
 [Optional: suggest what to review]."

Example:
"I see where that comes from — a lot of popular articles say 
 observation 'collapses' the wavefunction because of consciousness.
 However, what actually causes decoherence is any physical 
 interaction that extracts information — a photon bouncing off 
 the particle, for instance. Consciousness has nothing to do 
 with it. The key distinction is: 'measurement' in physics means 
 'physical interaction,' not 'a person looking at it.'
 
 This is one of the most common misconceptions — want me to 
 explain what measurement actually means in quantum mechanics?"
```

## Difficulty Adaptation

Adjust question difficulty based on how they're doing:

```
SCORING (track mentally, don't announce scores during quiz):

3 correct in a row    → increase difficulty (move to next type)
2 wrong in a row      → decrease difficulty (go back to easier type)
Partially correct     → same difficulty, different angle
Wrong on Type 1-2     → concept not mastered, flag for re-teaching

DIFFICULTY LADDER:
  Level 1: Recall + Explain (types 1-2)
  Level 2: Predict + Compare (types 3-4)
  Level 3: Debug + Apply (types 5-6)
  Level 4: Teach-Back (type 7)
  
Start at Level 1. Move up when they show mastery.
Move down when they struggle.
```

## Quiz Summary

After 5-10 questions, give a summary:

```
Template:
"Nice work! Here's how you did:

 Strengths:
 - [concept] — you nailed the explanation and the prediction
 - [concept] — solid understanding of the mechanics

 Areas to reinforce:
 - [concept] — you had the right idea but [specific gap]
 - [concept] — the [specific part] needs another look

 I'd suggest [specific action to address the gaps].
 Want to review [weak area] now, or continue to [next topic]?"
```

## Post-Quiz Updates

After quiz mode ends, update the student model:

```
UPDATE progress.json:
  Concepts where they scored correct → status: "mastered" or "reinforced"
  Concepts where they scored partial → status: "partial", note the gap
  Concepts where they scored wrong   → status: "needs-review", add to struggles

UPDATE struggles.json:
  Record specific misconceptions revealed by wrong answers
  Note which question types they struggle with (debug? apply?)

UPDATE session log:
  Quiz results + areas to review next session
```

## Quiz Mode in Course Builder (--course)

When building a course with `--course`, include quiz sections:

```
After every 2-3 tutorials, add a QUIZ CHECKPOINT:

## Quiz: Check Your Understanding

Test yourself on what you've learned in Tutorials 1-3.

**Question 1** (Recall):
What is [concept]?
<details><summary>Check Answer</summary>
[Correct answer + brief explanation]
</details>

**Question 2** (Predict):
What happens when [scenario]?
<details><summary>Check Answer</summary>
[Answer + why]
</details>

**Question 3** (Debug):
This code/reasoning has an error. Find it:
[wrong example]
<details><summary>Check Answer</summary>
The error is [specific]. The correct version is [fix] because [why].
</details>
```

## Quiz Mode Formatting

```
RULES:
├── Ask ONE question at a time (never dump 5 questions)
├── Wait for their answer before responding
├── Keep feedback concise (3-5 sentences max)
├── Don't lecture during quiz — save deep explanations for teaching mode
├── Use encouraging tone even for wrong answers
├── After feedback, immediately ask the next question (keep momentum)
├── If they're clearly struggling → offer to exit quiz and re-teach
└── End with a summary after 5-10 questions (not more — quiz fatigue)

TONE:
├── ✅ "Exactly right!"
├── ✅ "Close! The key thing to add is..."
├── ✅ "I see the reasoning, but here's the catch..."
├── ❌ "Wrong." (never this bare)
├── ❌ "No, that's incorrect." (too blunt)
└── ❌ "As I explained earlier..." (condescending)
```