A structured framework for extracting universal philosophical principles from concrete technical experiences. Built for use with Claude Code skills system.
| Layer | What it captures | Output |
|---|---|---|
| Layer 1: Breakthrough Path | Raw timeline of attempts, outcomes, decision points | Chronological list with metrics |
| Layer 2: Domain Knowledge | Field-specific insights, diagnostics, boundary conditions | Decision trees, comparison tables |
| Layer 3: Universal Principles | Cross-domain patterns that transfer across fields | Named principles with validation |
Each Layer 2 insight is systematically processed through:
- INVERSION - What is the opposite, and when would it be correct?
- GENERALIZATION - Strip domain terms, what is the abstract structure?
- TRANSFER - Where else does this pattern appear? (2+ domains)
- PARADOX - What contradiction does this resolve?
- META - What does this reveal about the problem-solving process?
Every Layer 3 principle must pass all six checks:
- Domain independence (no jargon)
- Predictive (would knowing this have changed your approach?)
- Multi-domain (2+ fields)
- Non-trivial (not obvious)
- Actionable (concrete change suggested)
- Falsifiable (evidence that would disprove it)
references/principles-from-competitions.md contains 11 validated universal principles extracted from two projects:
- Kaggle Store Sales (6 principles): Distribution Mismatch, Decoupling, Context-Dependent Tool, Workaround Trap, Diagnosis-First, Duality
- Knowledge System Redesign (5 principles): Meta-Knowledge Trap, Pruning-Over-Adding, External Validation, Maintenance Debt, Invisibility of Accumulation
- Reflexion (Shinn 2023)
- ExpeL (Zhao 2024)
- Common Wisdom Model (Grossmann 2020)
- Structure-Mapping Theory (Gentner 1983)
- Multi-Actor Insight Extraction (Nature Scientific Reports, 2025)
Copy the three-layer-wisdom-extraction/ directory to ~/.claude/skills/:
cp -r three-layer-wisdom-extraction ~/.claude/skills/- claudeception - Layer 2 domain knowledge extraction
- skill-refresh - Knowledge maintenance over time
MIT