I build Tierward - governed AI coding for teams on Claude Code. The principle is simple: Claude generates, your team decides.
You adopt it one tier at a time, starting with a light discovery setup and adding governance as you need it. The guardrails are mechanical: STOP gates before implementation, a set of audit skills, and a Stop hook that won't mark a task done until the tests pass. There's also an MCP server, so any AI-aware client can query a project's governance state directly.
- 🔭 Building Tierward • MIT-licensed, on npm as
tierward - 💬 Ask me about Claude Code, AI governance, and MCP
- 📍 Milan