← ClaudeAtlas

first-principles-explainlisted

Explain any concept, bug, or mechanism from first principles via Problem → Truths → Root Cause → Rebuild chain. Use when user asks "what is X", "how does X work", "explain X", "why use X", "why does this bug happen", "explain this code".
xshubhamg/voidskills · ★ 0 · AI & Automation · score 60
Install: claude install-skill xshubhamg/voidskills
# First-Principles Explain v2 Explain so the user gets the *why*, not just the *what*. Derive the concept as the inevitable fix to a real pain — never definition-first. ## Mode detector (run first, 1 line) - **Concept mode** (`what is / how does X work`): follow Chain below. - **Broken-code mode** (`why does this break / explain this snippet`): start from the pasted code, name what breaks observably, then follow same Chain. See `references/decomposition.md`. ## The Chain (no skipping) ### 1. Before: Life Without the Concept - What did we do before? Show 2 concrete pains with tiny example: "every re-render does X, so Y happens..." - Why it hurts: slow, buggy, confusing, wasteful. Make pain feel real. - For technical topics: show broken → fixed contrast with minimal code. ### 2. Surface Assumptions + Truths - List 2-3 hidden assumptions ("we assumed render is cheap / order doesn't matter"). - Mark each: Assumed (convention) vs True (physics/math/runtime guarantee). See `references/decomposition.md`. - Keep to a 3-row mini-table max in main answer. ### 3. Root Cause (one crisp line) - Format: "So the root cause is: we are [doing X when we don't need to / missing Y]..." - Go 1-2 levels deeper than obvious via 5-Whys (ask "why?" twice). This is the pivot — 1-2 sentences max. ### 4. The Pivot Question > "So how can we solve this problem?" Leave a beat before the answer. ### 5. Rebuild as Answer - Define [CONCEPT] in one simple sentence. Only now. - Show mechanism, not magi