# Harness²

> Harness the agent. Then harness its self-improvement.

Source: https://tools.launchllama.co/products/harness
Markdown: https://tools.launchllama.co/products/harness.md
Generated: 2026-11-01T14:22:25.694Z

## Facts

- **Name:** Harness²
- **Tagline:** Harness the agent. Then harness its self-improvement.
- **Short description:** Harness the agent. Then harness its self-improvement.
- **Website:** https://harness2.github.io/
- **Categories:** Open Source, Developer Tools, Artificial Intelligence, GitHub
- **Listing type:** product_hunt
- **Upvotes on Launch Llama:** 202
- **Listed since:** 2026-11-01
- **Featured (paid) listing:** no

## Full description

Harness² from Google Research is a plug-and-play recursive agent harnessing framework that lets LLM agents automatically evolve their execution harnesses (prompts, tools, memory) for open-world tasks—without ground-truth verifiers. Dual recursion: task-level Propose-Probe-Compose refinement + harnessing-level accumulation of reusable experience. Up to +11.5 points on professional benchmarks like WorkBuddy-Bench.

## Notes for agents

- Generated on request from live Launch Llama data, so it is always current.
- Human page: https://tools.launchllama.co/products/harness
- Site overview: https://tools.launchllama.co/llms.txt (long form: https://tools.launchllama.co/llms-full.txt)
- Every listing as markdown: https://tools.launchllama.co/catalog.md
- Search the catalog over MCP (https://tools.launchllama.co/mcp) or REST: https://tools.launchllama.co/developers