/validate

Validate a startup idea end-to-end: search KB for related docs, run ПОТОК analysis, pick a stack, generate a PRD.

MCP Tools (use if available)

If MCP tools are available, prefer them over CLI:

• kb_search(query, n_results) — search knowledge base for related docs
• project_info() — list active projects with stacks

If MCP tools are not available, fall back to Grep/Glob or CLI commands.

Steps

1. Parse the idea from $ARGUMENTS. If empty, ask the user what idea they want to validate.

2. Search for related knowledge: If MCP kb_search tool is available, use it directly:

   • kb_search(query="<idea keywords>", n_results=5) Otherwise search locally:
   • Grep for idea keywords in .md files across the project and knowledge base Summarize any related documents found (existing ideas, frameworks, opportunities).

3. Deep research (optional): Check if research.md exists for this idea (look in 3-opportunities/<project-name>/ or docs/).

   • If it exists: read it and use findings to inform ПОТОК analysis and PRD filling (competitors, pain points, market size).
   • If it does not exist: ask the user if they want to run deep research first. If yes, tell them to run /research <idea> and come back. If no, continue without it.

4. Alignment check: Look for core principles docs (search for manifest.md or use kb_search(query="manifest principles")). Verify the idea aligns with:

   • Privacy-first / offline-first?
   • One pain -> one feature -> launch?
   • AI as foundation?
   • Speed over perfection?

5. ПОТОК analysis: Walk the idea through 6 layers. For each layer, provide a brief assessment:

   • Layer 1 – Epistemological: Is this within the circle of competence? What assumptions are unproven?
   • Layer 2 – Temporal: What’s the time horizon? Is it Lindy-compliant?
   • Layer 3 – Action: What’s the minimum viable action? Second-order effects?
   • Layer 4 – Stakes: What’s the risk/reward asymmetry? Survivable downside?
   • Layer 5 – Social: Reputation impact? Network effects?
   • Layer 6 – Meta: Does this pass the mortality filter? Worth the finite time?

6. Stack selection: Ask the user which tech stack to use. If MCP project_info is available, show detected stacks from active projects as reference. Common stacks:

   • ios-swift — SwiftUI, CoreML, StoreKit 2
   • nextjs-supabase — Next.js, React, Tailwind, Supabase
   • python-api — FastAPI, SQLAlchemy, PostgreSQL
   • kotlin-android — Jetpack Compose, Room
   • cloudflare-workers — Hono, D1, edge-first
   • astro-static — Astro, Cloudflare Pages
   • python-ml — uv, Pydantic, ChromaDB, MLX

7. Generate PRD: Create a PRD document at 3-opportunities/<project-name>/prd.md (or docs/prd.md if not in solopreneur KB). Use a kebab-case project name derived from the idea. Include:

   • Problem: Based on the idea and ПОТОК analysis
   • Solution: Core feature set (keep it minimal — one pain, one feature)
   • Target Market: Who has this pain? (use research.md pain points if available)
   • Tech Stack: Selected stack with key packages
   • Architecture Principles: SOLID, DRY, KISS, schemas-first
   • Success Metrics: How to measure if it works

8. Output summary:

   • Idea name and one-liner
   • Opportunity score (0-10) based on the analysis
   • Key risk and key advantage
   • Path to generated PRD
   • Recommended next action