| name | blitz |
| description | Use when the user wants to learn a new topic quickly, understand a complex subject, study for an exam, prepare a learning plan, or use accelerated learning techniques. Trigger phrases: "learn fast", "blitz", "understand quickly", "study guide", "learning roadmap", "master [topic]", "spaced repetition", "active recall", "Feynman technique", "first principles". |
Blitz — Universal Accelerated Learning Protocol
A 7-phase interactive protocol for learning any topic with maximum efficiency. Based on validated learning science: Feynman Technique, First-Principles Thinking, Active Recall, and Spaced Repetition.
The user invokes: /blitz <topic> [--phase=N]
- Without
--phase, run phases 1→7 in sequence.
- With
--phase=3, jump directly to Phase 3 (Active Recall Training).
- With
--phase=all, do a full pass through all phases.
Run each phase one at a time. Complete the phase, get user feedback, then proceed.
Phase 1 — Feynman Technique (Understand First)
Goal: Establish foundational understanding by explaining in simple terms.
-
Teach the topic in simple language as if explaining to a complete beginner.
- Use analogies, avoid jargon, keep it conversational.
- Cover: What is it? Why does it matter? How does it work at a high level?
-
Ask the user to explain it back in their own words.
- Prompt: "Now explain this back to me in your own words. I'll point out gaps."
-
Diagnose gaps. After their explanation, identify:
- What they got right — reinforce it.
- What they got wrong or omitted — clarify simply.
- Areas where their explanation revealed confusion — re-teach those parts.
-
Confirm understanding before moving on.
Sign off: "Phase 1 complete. Ready for first-principles breakdown? (y/n)"
Phase 2 — First-Principles Breakdown
Goal: Decompose the topic into irreducible components so the user understands essential logic instead of memorizing surface facts.
-
Break it down:
- What are the fundamental building blocks?
- What axioms or invariants are true regardless of context?
- Strip away conventions, tools, and implementation details — what's the core essence?
-
Show the chain of reasoning — build up from first principles to the full topic, step by step.
-
Ask: "Does any link in this chain feel unclear or unmotivated?"
- Address each confusion by decomposing further.
-
Cross-check understanding — ask one re-framing question that forces the user to reason from first principles:
- "Based on these fundamentals, what do you think happens when X changes?"
Sign off: "Phase 2 complete. Ready for active recall training? (y/n)"
Phase 3 — Active Recall Training
Goal: Force retrieval of information without notes, then evaluate and correct.
-
Warn the user: "This is retrieval practice — answer without looking at notes. Being wrong is productive."
-
Ask 3–5 hard questions that force the user to retrieve and apply knowledge:
- Mix of: factual recall, application, synthesis, "what if" scenarios.
- Questions increase in difficulty.
-
For each answer:
- Praise what's correct.
- Correct errors with specific, constructive feedback.
- If partially correct, identify the missing piece and explain why it matters.
-
Final score summary:
- Areas of strength.
- Areas needing review (link these to Phase 2 first principles).
- Suggested re-study focus.
Sign off: "Phase 3 complete. Want a spaced repetition schedule? (y/n)"
Phase 4 — Spaced Repetition Table
Goal: Generate a concrete review schedule to move information to long-term memory.
-
Categorize the material into:
- Core concepts — must internalize permanently
- Supporting details — useful context, less critical
- Nitpicks / trivia — recall when needed, not priority
-
Build the table:
| Cadence | What to Review | Method |
|---|
| Daily (first 3 days) | Core concepts | Active recall (Phase 3 style) |
| Weekly (first month) | Core + supporting | Explain from memory + connect to first principles |
| Monthly (quarter 1) | Core concepts only | Teach someone else or write a summary |
| Quarterly (year 1) | Full map + any gaps | Diagnostic quiz (Phase 7 style) |
-
Customize: Adjust cadence based on user's deadline, depth goal, and prior knowledge.
-
Export suggestion: "Copy this table to a note app or flashcard system (Anki, RemNote, etc.)."
Sign off: "Phase 4 complete. Want a distilled knowledge summary? (y/n)"
Phase 5 — Knowledge Summary (Frameworks & Mental Models)
Goal: Distill the topic into clear frameworks, mental models, and simplified structures for easy retention and application.
-
Extract the core frameworks:
- What are the 2–3 mental models that capture the essence?
- Can the topic be visualized as a map, tree, flowchart, or spectrum?
-
Create a one-page cheat sheet (can be markdown, mermaid diagram, or conceptual map):
- Core principles (3–5 bullet points max)
- Relationships between concepts
- Common pitfalls / anti-patterns
-
Memorable hook: A one-sentence summary. "If you remember only one thing about this topic, it's _."
-
Ask: "Does this summary match your mental model? What would you add or change?"
Sign off: "Phase 5 complete. Want a learning roadmap for deeper mastery? (y/n)"
Phase 6 — Learning Roadmap (Acceleration Plan)
Goal: Create a staged plan for going from current level to mastery, with clear milestones.
-
Assess current level:
- Rate 1–10: How well does the user understand the topic after Phase 1–5?
- Identify the next logical growth area.
-
Build the roadmap:
| Stage | Goal | Activities | Checkpoint |
|---|
| Foundation | Solid understanding of basics | Review Phases 1–5, spaced repetition | Can explain core concepts without notes |
| Application | Use knowledge in real scenarios | Projects, exercises, real-world application | Completed 3 practical exercises |
| Integration | Connect to adjacent domains | Cross-domain reading, teach others | Created a guide or tutorial for others |
| Mastery | Push boundaries, contribute | Advanced resources, community participation, original work | Published analysis, contribution, or project |
-
Resources: Recommend 1–2 books, courses, or projects per stage. Not overwhelm — specific, actionable picks.
-
Time estimate: Realistic hours per stage based on topic complexity.
Sign off: "Phase 6 complete. Want a weak-point diagnosis to close remaining gaps? (y/n)"
Phase 7 — Weak Point Diagnosis
Goal: Identify remaining blind spots, misunderstandings, and suggest targeted fixes.
-
Ask 5 diagnostic questions — designed to expose:
- Common misconceptions about the topic
- Edge cases the user hasn't considered
- Connections between sub-topics they may have missed
- Practical application nuances
-
Analyze responses:
- Pattern: are errors in one specific area?
- Root cause: is the gap conceptual (didn't understand) or memory-based (forgot)?
-
Prescribe targeted improvements:
- For conceptual gaps → re-teach from first principles (Phase 2)
- For memory gaps → spaced repetition (Phase 4)
- For application gaps → practical exercise (Phase 6)
- For blind spots → specific reading or exercise
-
Final diagnosis summary:
Strengths: [what they know well]
Weak spots: [specific gaps]
Root causes: [why those gaps exist]
Rx: [targeted Phase references + one concrete next action]
Sign off: "Phase 7 complete. Run /blitz <topic> again in a week for reinforcement."
Quick Reference
| Phase | Name | Core Question | Duration |
|---|
| 1 | Feynman Technique | "Can you explain it simply?" | 5–10 min |
| 2 | First-Principles Breakdown | "What are the irreducible truths?" | 5–10 min |
| 3 | Active Recall Training | "Can you retrieve it under pressure?" | 10–15 min |
| 4 | Spaced Repetition Table | "What to review, when?" | 5 min |
| 5 | Knowledge Summary | "What's the one-page cheat sheet?" | 5–10 min |
| 6 | Learning Roadmap | "What's the path to mastery?" | 5–10 min |
| 7 | Weak Point Diagnosis | "What am I missing?" | 10 min |
Common Mistakes
- Skipping Phase 1 — "I already know this." Feynman reveals hidden gaps. Run it anyway.
- Rushing active recall — struggle is productive. Let the user sit with a hard question.
- Overloading the roadmap — recommend 1–2 resources per stage, not a book list.
- Too many diagnostic questions — 5 sharp questions > 20 shallow ones.
When NOT to Use
- The user needs a quick fact, not deep understanding (use search / direct answer).
- The topic is purely procedural / muscle memory (use practice drills, not this protocol).
- The user explicitly wants just one thing (references, a roadmap, etc.) — use
--phase=N.