Overview
MetaHarness is a tool that mints custom AI agent harnesses from any repository. It is described as a factory for agent frameworks, not another agent framework. The model is replaceable; the harness is the product. It generates a repo-aware CLI, a repo-aware coding agent, a local MCP server, memory scoped to the project, skills generated from the actual file layout, governance policy, release verification, and witness-signed provenance.
In under 60 seconds, in a browser, with nothing leaving the machine, it provides a custom AI agent harness for a repo, recommended agents, skills, slash commands, MCP tools, a scoped memory namespace and governance policy, and witness-signed provenance and release gates. The output drops into various agent hosts including Claude Code, OpenAI Codex, pi.dev, Hermes, OpenClaw, RVM, or Prime Agent. It is an npm-publishable .zip with the user's name, branding, and npx CLI.
Additional features include a score command to evaluate a repo before building, a router to pick the cheapest model that meets quality, and Darwin Mode for self-improvement. MCP is included as a first-class adapter surface with default-deny policy. It offers 19 quick-start vertical templates.
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Teams use MetaHarness to generate a custom AI agent harness for their repository.
- Teams use MetaHarness to score a repository before building a harness.
- Teams use MetaHarness to route requests to the cheapest model that meets quality.
- Teams use MetaHarness to let their harness improve itself via Darwin Mode.
- Teams use MetaHarness to publish their harness as an npm package for their organization.
- Teams use MetaHarness to run their harness on various agent hosts.
- Teams use MetaHarness to scan MCP tools for security issues.
Features
Scaffold focused custom agent harnesses
Give each harness its own branding
Generate a dedicated npx CLI
Include an MCP server
Add memory and a learning loop
Support witness-signed release workflows
Reviews
No reviews yet. Be the first to share your experience!