StackMap
Subscribe
Explore / harness2
google-research

harness2

Google Research's recursive harnessing framework: an agent refines its own harness (prompts, tools, memory, skills) per task via Propose–Probe–Compose, with model weights fixed.

22,822 2,277 Python Apache-2.0updated 4 days ago
agentsevals
View on GitHubDispute this mapping →
Curator's take

Use it if you research or tune agent harnesses and want the agent to improve its scaffolding on unseen tasks without ground-truth labels. Lessons carry forward between tasks. It is a research framework: expect paper-grade code and extra model calls per task for probing. Skip it if you want a ready coding agent or a production orchestration runtime.

Mapped by ShipWithAI editors · links verified

Continue your stack

What teams reach for next — and why each earns a place beside harness2. Ranked by curator confidence.

alternativealternativealternativeJITagentic-context-engineautoharnessharness2
pairs wellalternativebuilt withpick a node for the why · open it from the panel
Weekly digest
README.md1 min read

Harness²
Recursive Agent Harnessing for an Open World

Paper Website Benchmark guides License

Install, set up, and run JobBench: uv pip install -e '.[render]', harness2-setup jb, then harness2 --bench jb --k 1 --mode parallel. Click for prerequisites and commands.

A harness is the system around a model that enables it to act as an agent: prompts, tools, memory, skills, orchestration, and more. Harnessing is the process of optimizing that system.

Harness² is a plug-and-play framework for recursive agent harnessing in an open world, where agents encounter tasks unseen during harness design. It connects two levels of recursion while keeping model weights fixed:

  • Task-level recursion refines the execution harness for the current task through a contrastive Propose–Probe–Compose step: propose edits, probe their effects on the same task, and compose a refined harness from behavioral contrasts between executions, without ground-truth feedback.
  • Harnessing-level recursion updates a persistent improvement harness with adopted harnesses and evidence-supported lessons from each task to guide future refinement.

Harness² pipeline. Task-level recursion: an agent executor runs the initial harness, a harness editor proposes invariant and domain edits, candidate probes execute them, and a harness composer combines them into the final harness. Harnessing-level recursion: an experience reflector reflects on the step and updates the improvement harness used for the next task.

Across three professional benchmarks and four model–harness configurations, a single Harness² step improves GLM-5 and Gemini 3.5 Flash over their base harnesses, with gains of up to 11.5 points on WorkBuddy-Bench, 4.5 points on JobBench, and 11.8 percentage points on LAB's partial pass rate. Increasing the task-level recursion budget brings further gains.

The release supports parallel and sequential recursion, OpenCode and Codex executors, and JobBench, LAB, and WorkBuddy-Bench.

Video overview

https://github.com/user-attachments/assets/76b10548-b14b-4c8f-9aa5-a93662df4d56

Contents

Harness components

Harness² exposes eight editable components around the model: system prompt, guardrails, memory, tool guidance, subagents, skills, scripts, and plugins. They control context construction, the tool interface, delegation, and behavior inside the tool loop.