Awesome Harness Engineering

agent-sandbox enables easy management of isolated, stateful, singleton workloads, ideal for use cases like AI agent runtimes.

Amazon Bedrock Agentcore accelerates AI agents into production with the scale, reliability, and security, critical to real-world deployment.

Agent harness to make your slop code well-engineered and beautiful.

Inspect: A framework for large language model evaluations

Sub-millisecond VM sandboxes for AI agents via copy-on-write forking

Open-source, end-to-end platform for evaluating, observing, and improving LLM and AI agent applications. Tracing · Evals · Simulations · Datasets · Gateway · Guardrails. Self-hostable. Apache 2.0.

Build, Evaluate, and Deploy GUI Agents — online RL training, standardized benchmarks, and real-device deployment in one framework.

Give Claude Code a memory that evolves with your codebase. Hooks automatically capture sessions, the Claude Agent SDK extracts key decisions and lessons, and an LLM compiler organizes everything into structured, cross-referenced knowledge articles - inspired by Karpathy's LLM Knowledge Base architecture.

MCP server that gets Claude to 97.9% (188/192) on a real coding benchmark at -80% active tokens and -83% wall time, vs 78.3% plain. Structural code navigation + persistent memory engine. Works with every MCP client.

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.

Deterministic orchestrator for CLI coding agents (Claude Code, Codex, Gemini CLI, +40 more). No model in the coordination loop, so parallel runs in per-task git worktrees replay byte-identically. Signed lineage plus an opt-in HMAC audit chain a reviewer checks offline, without rerunning it. Cluster mode, air-gap deploy. https://bernstein.run

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.

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

Meta Harness Implementation

Continual harness optimization

swebench repro script for running confucius-code-agent (CCA)

No description.

ML6 x AISO Agent Workshop (March 2026)