Validate Idea
Test whether a business idea is worth pursuing — before writing a single line of code or spending a dollar.
Pure prompt skill: no file access, no network calls, no dependencies; minimal permission surface and transparent data flow. Deducted for: unknown repository license metadata and attribution that relies only on a README link without verifiable per-claim citations.
Instructions are self-consistent with a clear flow (problem → manual solution → payment validation → four questions → flags → verdict); text-only with no runtime dependencies; static cap of 10 applies. Deducted for: no abnormal-input handling, no failure-feedback conventions, and subjective verdict criteria that limit reproducibility.
Clear trigger (user has a business idea to validate) and audience; red/green flag lists give partial boundaries. Deducted for: no declared capability boundaries or non-fit ranges, no Chinese-language consideration; no environment dependency, so no reachability concern.
Well-structured with reasonable progression; README covers installation and a skill index; attribution is clear and the publisher matches the book's author, a checkable source. Deducted for: no versioning/changelog, no known-limitations disclosure, no example outputs or FAQ, missing license metadata.
As a structured advisory prompt it should guide users through validation and produce a three-way verdict, offering marginal value over an unstructured chat. Deducted for: static review cannot verify output quality, no representative output samples, and value depends heavily on model behavior and user participation.
Claims are tied to The Minimalist Entrepreneur with an official site link, and the publisher matches the book's author, giving a limited auditable source. Deducted for: no tests or executable reproduction, framework relayed without chapter-level citation, thin coverage; static cap of 5 applies.
- Static review only, nothing executed; verdict quality (Validated/Needs more validation/Pivot) cannot be guaranteed in advance — review conclusions manually.
- Repository license metadata is missing; verify licensing terms before commercial use or redistribution.
- Verdict criteria are subjective; identical inputs may yield different outputs; thresholds like '10 customers, 3 willing to pay' are heuristics, not norms.
- Content paraphrases a book without per-claim citations; for deeper application, consult the original book.
What does this skill do, and when should you use it?
This skill packages the idea-validation method from The Minimalist Entrepreneur by Sahil Lavingia into a Claude skill. Acting as a business advisor, it walks you through a four-step process: defining the problem, testing whether you can solve it manually first, checking whether people will actually pay, and answering four questions to ask before building. Its core principle is that validation happens through selling, not building. It ends with a clear verdict: validated, needs more validation, or pivot.
This is a pure prompt-based skill: it reads your described idea and evaluates it against the book's framework. It pushes you to name a precise target customer ('freelance graphic designers who struggle with invoicing', not 'businesses'), analyze existing workarounds, and rate pain intensity. It guides you toward solving the problem by hand for a few people first ('processizing', as Gumroad did when Sahil manually paid creators via PayPal one by one). It applies concrete evidence standards: talked to at least 10 potential customers, at least 3 willing to pay. It checks the book's four pre-build questions (shippable in a weekend, makes life a little better, someone will pay, fast feedback loop), screens for red flags (nobody has this problem, only 'my friends think it's cool') and green flags (people already paying for inferior solutions, scratching your own itch), then delivers one of three verdicts: Validated, Needs more validation, or Pivot. It calls no scripts, tools, or network requests.
- An aspiring side-project founder with an idea who is unsure whether to invest build time
- A developer weighing quitting a job to build a SaaS, wanting to test willingness to pay first
- A founder halfway through building a product nobody is buying, who wants to backfill demand validation
- A member of a specific community (designers, teachers, etc.) wanting to solve a problem the community complains about daily
- A reader of The Minimalist Entrepreneur who wants to apply the book's validation checklist in practice
What are this skill's strengths and limitations?
- Complete and concrete methodology: four steps, four pre-build questions, and red/green flag checklists are actionable criteria, not vague advice
- Forces an explicit three-way verdict (validated / needs more validation / pivot) instead of hedging
- Grounded in real book examples (Gumroad's manual PayPal payouts) that make processizing tangible
- Pure prompt implementation with zero runtime dependencies — works on any platform that supports skills or system prompts
- Judgments rely entirely on information you supply in conversation; it cannot independently verify market data or whether customer interviews actually happened
- Output quality tracks input quality — vague descriptions of the idea and customer yield generic results
- No license declared in the repo, and no tests or sample conversations provided; real-world effectiveness beyond the author's claims is undocumented
- Covers validation only; MVP building, customer acquisition, and pricing live in sibling skills in the same repo and should not be attributed to this one
How do you install this skill?
In Claude Code run:
/plugin marketplace add slavingia/skills
/plugin install minimalist-entrepreneur
This installs the whole collection (10 skills including this one). Alternatively clone locally: git clone https://github.com/slavingia/skills.git ~/.claude/plugins/skills, then run the same /plugin commands. The README does not document installing a single skill on its own, nor installation on other platforms such as Codex or ChatGPT.
How do you use this skill?
After installation, invoke it in Claude Code via /validate-idea, or simply describe your idea and say you want to test it before building (its description is written so the model can auto-trigger). Example opener: 'I have an idea for contract management for freelance photographers — help me validate whether it's worth pursuing.' Answer its questions step by step; you'll end with one of three verdicts plus next steps.