best-of-agent-harnesses
byryanalberts· 1 MCP server
A model answers; an agent acts. An agent harness is the runtime that turns one into the other — the model thinks; the harness decides what that thinking is allowed to touch.
How to install best-of-agent-harnesses?
ryanalberts/best-of-agent-harnesses
$git clone https://github.com/ryanalberts/best-of-agent-harnessesInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Install best-of-agent-harnesses by running `git clone https://github.com/ryanalberts/best-of-agent-harnesses`, then use it for the current task and follow its documentation at https://github.com/ryanalberts/best-of-agent-harnesses.
README.md
| 1 | <!-- markdownlint-disable --> |
| 2 | <h1 align="center"> |
| 3 | Best of Agent Harnesses and Harness Techniques |
| 4 | <br> |
| 5 | </h1> |
| 6 | |
| 7 | <p align="center"> |
| 8 | <strong>🏆 Curated list of AI agent harnesses, orchestration frameworks, and harness techniques for reliable agentic systems.</strong> |
| 9 | </p> |
| 10 | |
| 11 | <p align="center"> |
| 12 | <a href="https://best-of.org" title="Best-of Badge"><img src="http://bit.ly/3o3EHNN"></a> |
| 13 | <a href="#contents" title="Project Count"><img src="https://img.shields.io/badge/projects-135-blue.svg?color=5ac4bf"></a> |
| 14 | <a href="https://ryanalberts.github.io/best-of-Agent-Harnesses/" title="Browse the searchable site"><img src="https://img.shields.io/badge/website-live-5ac4bf.svg"></a> |
| 15 | <a href="#for-agents" title="Agents can query this list — MCP server, llms.txt & JSON"><img src="https://img.shields.io/badge/agents-query%20this%20list-5ac4bf.svg"></a> |
| 16 | <a href="#contribution" title="Contributions welcome"><img src="https://img.shields.io/badge/contributions-welcome-green.svg"></a> |
| 17 | <a href="https://github.com/RyanAlberts/best-of-Agent-Harnesses/commits/main" title="Updates"><img src="https://img.shields.io/github/last-commit/RyanAlberts/best-of-Agent-Harnesses?color=green&label=updated"></a> |
| 18 | </p> |
| 19 | |
| 20 | <p align="center"> |
| 21 | 🌐 <strong><a href="https://ryanalberts.github.io/best-of-Agent-Harnesses/">Browse the searchable site</a></strong> — one page per harness, filter by capability, autonomy & recovery. |
| 22 | </p> |
| 23 | |
| 24 | <p align="center"> |
| 25 | 🤖 <strong>Agents can query this list</strong> — an <a href="#for-agents">MCP server</a> (<code>recommend</code>, <code>pick_harness</code>, …), <a href="llms.txt">llms.txt</a> & <a href="harnesses.json">JSON</a>, so your agent recommends harnesses too. |
| 26 | </p> |
| 27 | |
| 28 | <p align="center"> |
| 29 | 🧡 <strong>A curated list is only as good as the people who stop mid-scroll to point at what it's missing.</strong><br> |
| 30 | These folks did exactly that — found a gap, wrote it up, and made the list better than one maintainer ever could. |
| 31 | <a href="#-thank-you-contributors"><strong>Meet the 7 →</strong></a> |
| 32 | </p> |
| 33 | |
| 34 | ## What is an agent harness? |
| 35 | |
| 36 | A model answers; an agent acts. An agent harness is the runtime that turns one into the other — the model thinks; the harness decides what that thinking is allowed to touch. |
| 37 | |
| 38 | Every prior wave of automation was constrained by brittleness: you scripted exact behavior, and when the world deviated, the system broke. Foundation models inverted that problem—they're flexible but directionless, stateless, and disconnected from anything real. The agent harness exists to bridge that gap: it is the orchestration infrastructure that converts a model's per-turn reasoning into sustained, tool-using, error-recovering, goal-directed behavior across time. Architecturally, it plays the role the kernel played in operating systems or the controller played in industrial robotics—mediating between raw capability and a messy environment—but with a critical difference: the "capability" it governs is general-purpose cognition, which means the harness is simultaneously a scheduler, a permission system, a memory manager, and a policy enforcement layer, all under-specified and evolving in real time. |
| 39 | |
| 40 | ## Why harnesses matter |
| 41 | |
| 42 | Better models make harnesses more important: more capabilities mean more failure modes, and production needs retry logic, fallbacks, and validation. Harness quality—not just model quality—determines whether agents actually ship. This list ranks projects by relevance to harness concerns (environment, orchestration, lifecycle, guardrails) and by stars/activity. |
| 43 | |
| 44 | ## The landscape at a glance |
| 45 | |
| 46 | [](assets/landscape.svg) |
| 47 | |
| 48 | _Every project in the list, plotted by adoption surface area (the [simplicity ↔ capability axis](#guide-to-rankings)) against GitHub stars. Colors are categories; the largest projects in each tier are labeled._ |
| 49 | |
| 50 | [
