Google Research 开源 Harness²：面向开放世界的递归 Agent Harness 框架

Recursive Agent Harnessing for an Open World 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. 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 - [Video overview](#video-overview) - [Harness components](#harness-components) - [Install](#install) - [Run a benchmark](#run-a-benchmark) - [Experiments](#experiments) - [Web access](#web-access) - [Results and resume](#results-and-resume) - [Citation](#citation) - [Contributing](#contributing) - [License](#license) - [Disclaimer](#disclaimer) ## 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. ## Install Use **Linux, Python 3.12, and [uv](https://docs.astral.sh/uv/getting-started/installation/)**. From the repository root: ```bash uv venv --python 3.12 .venv source .venv/bin/activate uv pip install -e '.[render]' harness2-demo --output runs/demo --k 2 ``` The local demo runs both recursion modes with deterministic editors and real Python executions. It requires no model credentials. For model-backed benchmarks, install **Node.js 20+, the Google Cloud CLI**, and the [benchmark dependencies](docs/SETUP.md#requirements). LAB needs Pandoc and Poppler; WorkBuddy-Bench needs Docker. Then configure your models: ```bash cp .env.example .env # Edit .env with your Vertex AI project and model IDs. gcloud auth application-default login ``` See the [setup guide](docs/SETUP.md) for model settings, authentication alternatives, and the [Codex executor](docs/SETUP.md#codex-executor). ## Run a benchmark The following example installs LAB and runs one task, then evaluates and reports its result. Model-backed runs incur API costs; start with a small task selection. ```bash harness2-setup lab harness2-check --bench lab harness2 --bench lab --k 1 --mode parallel \ --only-domains antitrust-competition --max-tasks 1 harness2-judge --bench lab --k 1 --mode parallel harness2-report --bench lab --k 1 --mode parallel ``` Sequential recursion runs the same commands with `--mode sequential`. `--k` is the task-level recursion budget; see [Experiments](#experiments). Judging and reporting must use the same benchmark, executor, subset, `--k`, `--mode`, and seed as the refinement run. Add `--dry-run` to `harness2` to preview task selection without model calls. | Benchmark | Tasks | Install | Guide | |---|---|---|---| | JobBench | Professional work | `harness2-setup jb` | [JobBench](docs/BENCHMARKS.md#jobbench) | | LAB | Legal work | `harness2-setup lab` | [LAB](docs/BENCHMARKS.md#lab) | | WorkBuddy-Bench | Office, web, and code | `harness2-setup wb` | [WorkBuddy](docs/BENCHMARKS.md#workbuddy-bench) | Choose one benchmark per command. WorkBuddy also takes `--subset office`, `--subset web`, or `--subset code` on every command after setup. The [benchmark guides](docs/BENCHMARKS.md) cover each one's data, domains, and grading. ## Experiments Scale task-level recursion in two ways: **parallel recursion** explores multiple candidate harnesses and composes refinements from their execution evidence; **sequential recursion** carries the refined harness and accumulated evidence through successive Propose–Probe–Compose steps. Parallel recursion explores wider; sequential recursion refines deeper. | Option | Behavior | |---|---| | `--k 1` | One Propose–Probe–Compose step | | `--k K --mode parallel` | Probe `K` general and `K` domain-specific candidates, then produce `K` compositions | | `--k K --mode sequential` | Refine the harness through `K` successive steps | | `--substrate cx` | Use Codex instead of the default OpenCode executor | After each task, harnessing-level recursion retains the adopted harness and evidence-supported lessons to guide future refinement. See the [experiment guide](docs/EXPERIMENTS.md) for task selection, concurrency, reproducibility, and baseline comparisons. ## Web access Executor network access is unrestricted, and Codex runs with its sandbox disabled. Public benchmark rubrics may be accessible online. Before evaluation, configure [network restrictions](docs/EXPERIMENTS.md#web-access) to prevent access to benchmark answers and grading criteria. ## Results and resume Results are saved under `runs/`, organized by benchmark and subset where applicable. They include task summaries, deliverables, adopted harnesses, reflection notes, and evaluation results. Repeat the same command to resume. Use a new `HARNESS2_RUNS_DIR` when changing models or task selection. See [results and resume](docs/EXPERIMENTS.md#results-and-resume) for details; each command also supports `--help`. ## Citation ```bibtex @article{xu2026harness2, title = {Harness$^2$: Recursive Agent Harnessing for an Open World}, author = {Xu, Ruiyao and Chen, Yanfei and CuiZhu, Zhongying and Dalvi Mishra, Bhavana and Ming, Yifei and Yu, Han and Han, Rujun and Lee, Chen-Yu and Pfister, Tomas}, journal = {arXiv preprint arXiv:2610.14224}, year = {2026}, url = {https://arxiv.org/abs/2610.14224} } ``` ## Contributing See [CONTRIBUTING.md](CONTRIBUTING.md). ## License Harness² is licensed under [Apache 2.0](LICENSE). Third-party code under [`third_party/`](third_party/README.md), benchmarks, and datasets retain their upstream licenses; see [third-party notices](dependencies/README.md). ## Disclaimer This is not an officially supported Google product. This project is not eligible for the [Google Open Source Software Vulnerability Rewards Program](https://bughunters.google.com/open-source-security).