Awesome Agent Harness

Official AHE code — Agentic Harness Engineering: observability-driven automatic evolution of coding-agent harnesses (concurrent w/ meta-harness). NexAU-AHE reaches 84.7% ± 2.1 pass@1 on Terminal-Bench 2 (GPT-5.5). Lifts GPT-5.4 69.7→77.0% over 10 iters, beats Codex/ACE/Training-Free GRPO; frozen harness transfers to SWE-bench-Verified.

CUGA is an open-source generalist agent harness for the enterprise, supporting complex task execution on web and APIs, OpenAPI/MCP integrations, composable architecture, reasoning modes, and policy-aware features.

No description.

A fully customizable and self-hosted sandboxing solution for AI agent code execution and computer use. It features out-of-the-box support for backtracking, a simple REST API and Python SDK, automatic port forwarding, and secure MicroVM isolation. Perfect for safely running, testing, and backtracking multi-step agent workflows.

Security Governance for Agentic AI

A production-ready runtime framework for agent apps with secure tool sandboxing, Agent-as-a-Service APIs, scalable deployment, full-stack observability, and broad framework compatibility.

Open-source cross-agent memory layer for coding agents via MCP. Compatible with Claude Code, Codex, Cursor, Windsurf, Gemini CLI, Antigravity, OpenClaw, Hermes Agent, Oh-my-Pi, Pi, Copilot, Kiro, OpenCode, and Trae.

Ultimate Context Engineering Infrastructure, starting from MCPs and Integrations

An agent benchmark with tasks in a simulated software company.

Claw-Eval is an evaluation harness for evaluating LLM as agents. All tasks verified by humans.

CheetahClaws: A Fast and Easy-to-Use Agent Harness Infrastructure for Long-Horizon, Multi-Model, and Tool-Using AI Systems

Collection of evals for Inspect AI

Agent harness and tookit for building AI agents and agentic applications. CLI and SDKs included

Measuring and evolving with the frontier of agent work

Sandboxed code execution for AI agents, locally or on the cloud. Massively parallel, easy to extend. Powering SWE-agent and more.

🤖 MateClaw — Your second brain with Multi-Agent Orchestration, MCP Protocol, Skills & Memory, Dream, and Multi-Channel Support. Built on Spring AI Alibaba.

Bring your own agent and build a self-improving agentic system. Automatically mine failures, optimize the agent harness, and gate against regressions.

SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?

An in-the-wild benchmark for AI agents in the production harness.

The action firewall for AI agents. Enforce policy and human approval before risky tool calls, shell commands, workflows, and production changes, with auditable evidence.

Self-hosted Personal AI + agent runtime in .NET (NativeAOT-friendly)

Safe runtime for autonomous on-chain AI agents: isolated sandboxes, Library skills, encrypted secrets.

Deterministic safety solutions for probabilistic AI agents

A collection of Model Context Protocol (MCP) servers, clients and developer tools by IBM.

A generative AI-powered framework for testing virtual agents.

The production-ready agent harness framework for Python

Remote approvals, policy checks, and execution evidence for unattended AI agents.

Secure runtime to sandbox AI agent tasks. Run untrusted code in isolated WebAssembly environments.

WorkArena: How Capable are Web Agents at Solving Common Knowledge Work Tasks?

Open Python agent harness for production AI apps: tools, MCP, memory, workspace, telemetry, subagents, background tasks, and OmniServe APIs.

Secure local dev environment for collaboration with AI coding agents

The harness layer for Claude Code — a reference implementation of harness engineering with hook-enforced dual review, state-machine gates that survive context compaction, and fail-closed safety where it counts. Quality gates that AI can't skip.

Universally Triggered Agent Harness - An OpenClaw-like Inngest-powered personal agent

Runtime for long-horizon agents

HexAgent – An Agent harness that gives any LLM a computer to complete tasks the way humans do

Evaluation harness for OpenHands V1.

Tandem is the authority layer between AI agents and company tools: runtime policy for tools, memory, approvals, and audit trails.

This repository defines AGENT.md, a standardized format that lets your codebase speak directly to any agentic coding tool.

No description.

Benchmark for comparing agent harnesses on everyday online tasks — fixes the base model, varies the harness. Sister project of ClawBench, same scoring pipeline.

A verified version of the WebArena Benchmark