Home
All projects

Agentic Meta-Harness

A meta-harness that wraps an AI coding agent and enforces eight deterministic, fail-closed engineering gates (build, hygiene, layout, format, lint, typecheck, tests, security) via Claude Code hooks. v0 governs Python projects, with a per-language/per-harness seam designed to extend to other stacks and agents.

Claude CodeCI GatesADRsPythonAgentic
Source
  • Borromean rings — three links that hold only together: remove any one gate and the whole guarantee falls apart
Borromean rings — three links that hold only together: remove any one gate and the whole guarantee falls apart

A meta-harness that wraps an AI coding agent and enforces eight deterministic gates — build, hygiene, layout, format, lint, typecheck, tests, and security — through Claude Code hooks. Today (v0) it governs Python projects, with a per-language / per-harness seam (a policy-spine TOML plus a uniform check contract) designed to extend to other stacks and agents.

The pipeline is fail-closed: it blocks an agent from “finishing” until every check passes, with the verifier kept external to the generator. Decisions are grounded in documented architecture-decision records.