← ClaudeAtlas

teardown-idealisted

Runs a brutal, evidence-first teardown of a startup/app idea in one of two modes — solo-dev side-income screen (default: build effort, Play Store/ASO crowding, realistic solo income, ship-it-or-skip verdict) or VC-scale investment teardown (market sizing, moats, unit economics, funding verdict) — producing a full research report either way. Use whenever the user pastes a startup/product/app idea and asks to evaluate, roast, stress-test, or tear it apart, tell them if it's good, do a pre-mortem, or asks "should I build this" / "is this worth pursuing" / "is this a good side-income idea" — even without the word "teardown". Also trigger for one-line ideas dropped mid-conversation if the user is clearly asking for real evaluation, not brainstorming. Do NOT use for open-ended ideation or for a PRD on something already committed — use a general brainstorming or spec-writing tool instead. Assumes the idea is bad until evidence says otherwise — never use if the user wants encouragement or a first-draft pitch.
sakshi-gupta-1/idea-teardown-skill · ★ 1 · Code & Development · score 64
Install: claude install-skill sakshi-gupta-1/idea-teardown-skill
# Startup / App Idea Teardown ## Why this exists Most startup evaluations are exercises in confirmation bias — the founder wants validation, and the evaluator (human or model) supplies it. This skill exists to invert that default. Treat the idea as guilty until the evidence proves otherwise. Brutal honesty beats encouragement here — honor that even when the idea sounds good, and especially when it's the user's own idea. The report is only as good as the evidence behind it. A polished-looking report built on assumptions dressed up as facts is worse than a rough one that's honest about what's unknown — it will get someone to make a bad call with false confidence. So the non-negotiable discipline running through every section, in both modes, is: **label everything as Fact, Evidence, Assumption, or Unknown, and never let assumptions masquerade as either of the first two.** Requires WebSearch. Uses Claude in Chrome (`mcp__claude-in-chrome__*`, load via ToolSearch if deferred) for Play Store / App Store review mining and listing checks when available — degrade gracefully to web search if it's not connected, and say so in the report rather than fabricating review counts or ratings. ## Step 0: Which lens — side-income or VC-scale? These are different questions with different pass criteria, not the same report with a lighter tone. A crowded market with weak differentiation kills a VC pitch; the same market can still be a perfectly fine solo side-income wedge if the build is chea