Generate Build-Ready Product Plans from Ideas

idea-os is a five-phase sequential pipeline that turns a raw problem statement into four build-ready files - clarifying questions, deep research, a PRD, and a phased execution plan with platform and stack picks, a user-journey diagram, and kill criteria. Each phase's output feeds the next: research shapes the PRD, the PRD shapes the plan, and the plan's kill criteria tie back to research insights. Unlike single-command PRD generators, it refuses to write a PRD until research is done, and refuses to write a plan until the PRD is stable - and its depth and vocabulary scale to a two-axis classification of idea complexity and builder sophistication, so a first-time founder isn't drowning in jargon while an experienced one gets full rigor.

Use it when a user shares a raw product idea or problem statement and wants a structured pipeline all the way from clarifying questions through research, a PRD, and an execution plan - files to take forward, not a one-shot answer - especially for a non-technical founder, PM, or hobbyist who needs structure bridging "idea" and an actual build queue. It is not for quick sanity-check feedback on a half-formed idea (that's idea-refine) or for editing an already-existing PRD (that's product-management); it is also better suited to new-idea validation and early planning than to late-stage optimization of a shipped product.

Input: a raw problem statement, classified in Phase 1 on two axes before anything else - idea tier (T1 weekend utility, T2 SaaS MVP or AI wrapper, T3 marketplace/B2B SaaS/regulated) and builder sophistication (S1 non-technical, S2 hobbyist, S3 founder or senior PM) - stated as a one-line classification like "T2 - S2" before proceeding. Output: questions.md (4-18 actionable questions grouped into who-and-pain, scope-and-wedge, and constraints-and-goals, with generic questions rejected, and execution paused until they're answered or autonomous-mode assumptions are declared); research.md (built from a minimum of 5 WebSearches and 2 WebFetches on named competitors, one source per TAM number, every source dated, anything unsourced flagged [assumption], covering problem validation, JTBD, TAM/SAM/SOM, competitor positioning, SWOT, distribution fit, risks, and 3-7 non-obvious insights); PRD.md (a falsifiable problem statement, named personas, ranked JTBD, mandatory non-goals, and leading/lagging metrics); and plan.md (a user journey in text and mermaid, a platform recommendation tied back to the research, a stack matrix across conservative/modern/cutting-edge options, a phased MVP-to-v1-to-target build with kill criteria and distribution per phase, and 3-5 immediate next actions).

Hands off to idea-refine for quick feedback on a half-formed idea instead of the full pipeline, and to product-management for editing a PRD that already exists; the upstream project (Slashworks-biz/idea-os, MIT-licensed) ships 11 reference files, 4 asset templates, and a 590-line worked example that this listing is a reference copy of.

Non-technical founders, PMs, and hobbyists who have a raw idea and want a structured, research-grounded path to a PRD and phased execution plan - rather than a quick one-shot answer - with the depth and vocabulary matched to how technical they are and how complex the idea is.