Back to directory

Harmonist

A multi-agent orchestration framework that mechanically enforces engineering rules on AI coding assistants, gating every code-changing turn on checks that reviewers ran, memory updated, and shipped files are supply-chain intact.

Harmonist

Harmonist is a drop-in multi-agent framework for Cursor, Claude Code, GitHub Copilot, Windsurf, Aider and other AI coding assistants that enforces engineering rules as a mechanical gate rather than a prompt-level request. Every code-changing turn is checked by hooks that verify whether required reviewer agents ran, whether project memory was updated, and whether the supply chain of every shipped file is intact; if a check fails, the turn does not complete no matter what the model claims.

It ships with a large catalogue of specialized agents and structured, validated memory so context persists between sessions, aiming to make rules such as running QA before merging or requiring security review before touching authentication code unable to be silently skipped, unlike frameworks that rely on the model to follow prompted instructions.

It has no runtime dependencies, is written in Python, and is released under the MIT license.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.