CI/CD for technical decisions.
You're about to mass-migrate 200 microservices. Slow down.
The migration will take months. It will cost real money. And right now, the decision to move is based on a Slack thread, a blog post, and a gut feeling from someone who left the company.
You'd never ship code without tests. Why ship a decision without validated evidence?
npx @grainulation/wheat "Should we migrate to GraphQL?"One command. Zero prompts. Sprint ready in under 3 seconds.
Then open your AI coding tool and start investigating:
/research "GraphQL performance vs REST"
/challenge r003
/blind-spot
/brief
Works with Claude Code, Cursor, GitHub Copilot, or standalone via CLI.
For native tool access in Claude Code:
claude mcp add wheat -- npx -y @grainulation/wheat mcpThis gives Claude direct access to wheat's claims engine — add-claim, compile, search, status — without shelling out.
npx @grainulation/wheat quickstartCreates a demo sprint with pre-seeded claims, an intentional conflict, compiles everything, and opens a dashboard. The compiler flags the conflict and blocks output until it's resolved.
Wheat is a continuous planning pipeline. Findings are validated as they come in:
You investigate → Claims accumulate → Compiler validates → Brief compiles
/research typed, evidence-graded 7-pass pipeline backed by evidence
/prototype
/challenge
Claim types: constraint, factual, estimate, risk, recommendation, feedback
Evidence tiers: stated → web → documented → tested → production
The compiler catches conflicts, flags weak evidence, and blocks the build when issues exist. You can't ship a brief built on contradictions — same as you can't merge with failing tests.
| Command | What it does |
|---|---|
/research <topic> |
Deep dive on a topic, creates claims |
/prototype |
Build something testable |
/challenge <id> |
Adversarial stress-test of a claim |
/witness <id> <url> |
External corroboration |
/blind-spot |
Find gaps in your investigation |
/brief |
Compile the decision document |
/status |
Sprint dashboard |
/present |
Generate a stakeholder presentation |
/resolve |
Adjudicate conflicts between claims |
Wheat installs two optional guard mechanisms:
- Git pre-commit hook — prevents committing broken claims
- Claude Code guard hook — prevents generating output from stale compilations
Wheat doesn't care what language you use or what AI tool you run. Your Scala project, your Python monorepo, your Flutter app — wheat works the same everywhere. Node 20+ is the only requirement. Zero npm dependencies.
| Tool | Role |
|---|---|
| wheat | Research engine — grow structured evidence |
| farmer | Permission dashboard — approve AI actions in real time |
| barn | Shared tools — templates, validators, sprint detection |
| mill | Format conversion — export to PDF, CSV, slides |
| silo | Knowledge storage — reusable claim libraries |
| harvest | Analytics — cross-sprint patterns and prediction scoring |
| orchard | Orchestration — multi-sprint coordination |
| grainulation | Unified CLI — single entry point to the ecosystem |
You don't need all eight. Start with wheat. That's it.
MIT
