The knowledge base behind oddkit — an open-source MCP server that gives your AI structured memory and epistemic discipline.
This repo is the fuel. oddkit is the engine. oddkit reads the markdown files in this repository and makes them available to your AI through structured tools. You can also point oddkit at your own repo to build your own knowledge base.
Step 1: Connect oddkit to your AI tool. See the oddkit repo for setup instructions — it takes 30 seconds.
Step 2: Read Getting Started with ODD and oddkit for the full walkthrough: connecting, trying it, bootstrapping your project, and building your own knowledge base.
This is a living knowledge base with 400+ documents spanning governance, methodology, planning, and public essays. It's organized into four tiers:
| Directory | What It Contains |
|---|---|
canon/ |
Governance — axioms, constraints, values, diagnostics, methods. The foundational principles that oddkit enforces. |
odd/ |
Methodology — ODD (Outcomes-Driven Development) philosophy, epochs, maturity model, getting-started guides. |
docs/ |
Implementation — planning documents, decision records, incident reports, tool documentation, session journals. |
writings/ |
Public essays — articles published on klappy.dev about AI-augmented workflows, knowledge transfer, and building systems that build systems. |
If you want to understand the philosophy:
If you want to understand the system:
If you want to build your own:
oddkit reads markdown files from any GitHub repo. You can point it at yours:
canon_url: "https://raw.githubusercontent.com/YOUR_ORG/YOUR_REPO/main"
Start with a few markdown files — decisions, constraints, learnings — and grow from there. oddkit reads what you write and makes it available to your AI. No schema required, no methodology to adopt. Start with what hurts.
For the full guide, see Getting Started with ODD and oddkit.
Built by Klappy — a systems architect with ~15 years in Bible translation technology, building systems that build systems.
oddkit repo: klappy/oddkit Website: klappy.dev
MIT