--- license: apache-2.0 pretty_name: Doc2Harness language: - en task_categories: - text-generation tags: - code - agents - coding-agents - harness-engineering - benchmark - llm-evaluation size_categories: - n<1K configs: - config_name: scenarios default: true data_files: - split: test path: data/scenarios.jsonl - config_name: leaderboard data_files: - split: test path: data/leaderboard.jsonl ---

Doc2Harness

Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents

🌐 Project page • 📄 Paper • 💻 Code

Doc2Harness measures how well a model can **build a working coding agent harness** from a development document, rather than how well it writes a patch inside an existing one. - **Stage 1, build.** A build model, working through a fixed builder harness (OpenCode) in an isolated container, receives a development document and an interface-only contract skeleton, and implements a complete candidate harness. - **Stage 2, evaluate.** The frozen candidate harness runs in containers separate from the construction environment, one per test case, and is driven end to end by a scripted **mock model**. Each test case is judged on four observable channels (requests sent to the model, files on disk, process exit code, command output), and **positive controls** rule out vacuous passes.

Doc2Harness pipeline

> [!NOTE] > **This is a partial release.** The full benchmark in the paper has **328 test > cases**. This dataset releases the same public subset as the > [GitHub repository](https://github.com/YuyaoGe/doc2harness): **94 scenarios**, > i.e. 53 scored test cases plus the 41 positive controls that gate them. The > remaining test cases are held out. The leaderboard numbers come from the full > suite and cannot be reproduced from this subset alone. > [!IMPORTANT] > Doc2Harness is an **executable** benchmark. The `scenarios` and `leaderboard` > configs are for browsing and analysis; `load_dataset` alone does not score > anything. To score a harness, use the full evaluation code shipped in > [`code/`](#run-the-benchmark) (a snapshot of the > [GitHub repository](https://github.com/YuyaoGe/doc2harness)) with Docker. ## Contents | Path | What it is | |---|---| | `data/scenarios.jsonl` | The 94 public scenarios, one per row (config `scenarios`) | | `data/leaderboard.jsonl` | Pass rates of 19 frontier models on the full 328-case suite (config `leaderboard`) | | `development_document/functional_spec_v6.md` | The development document given to the build model in Stage 1 | | `contract_skeleton/` | The interface-only contract skeleton (CLI, headless entry point, evaluation API), plus `AGENT_CONTRACT-v2.md` | | `code/` | Full evaluation code: Stage 1 build pipeline and Stage 2 framework (catalog, drivers, mock gateway), identical to GitHub commit [`5411f53`](https://github.com/YuyaoGe/doc2harness/tree/5411f53d497a6c76c1dd41bd55f08cc033203d2b) | ## Usage ```python from datasets import load_dataset scenarios = load_dataset("YuyaoGe/doc2harness", "scenarios", split="test") leaderboard = load_dataset("YuyaoGe/doc2harness", "leaderboard", split="test") scored = scenarios.filter(lambda r: r["scored"]) print(len(scenarios), len(scored)) # 94 53 ``` The development document and the skeleton are plain files: ```python from huggingface_hub import snapshot_download path = snapshot_download("YuyaoGe/doc2harness", repo_type="dataset", allow_patterns=["development_document/*", "contract_skeleton/**"]) ``` ## Run the benchmark **Requirements:** Python 3.11+ and Docker. Both stages use the standard library only. ```bash pip install -U huggingface_hub hf download YuyaoGe/doc2harness --repo-type dataset --include "code/*" --local-dir doc2harness cd doc2harness/code # sanity check: loading the catalog verifies its digests PYTHONPATH=src python3 -c "from doc2harness.benchmark.dispatch import active_scenario_ids as a; print(len(a()))" # 94 ``` **Stage 2 only: score an existing harness.** No model key is needed; scenarios drive the product against a mock gateway. ```bash docker build --target evaluation -t doc2harness/evaluation:candidate -f containers/Dockerfile . PYTHONPATH=src:tools:run python3 tools/run_production_catalog.py \ --artifact /path/to/product --artifact-id my-product \ --output /tmp/scores --concurrency 12 --timeout 300 ``` Each scenario writes a JSON receipt, plus a `catalog_receipt.json`. The score is `pass / (pass + fail)` over applicable scenarios; a scenario that cannot be observed on a product is `INCONCLUSIVE`, not a failure. Narrow the selection with `--family headless` or `--scenario-id `. **Stage 1: build a new harness, then score it.** This needs the builder image and a local LiteLLM gateway holding your own API key; see `code/src/stage1/README.md`. ```bash docker pull ghcr.io/yuyaoge/doc2harness-builder:latest PYTHONPATH=src python3 src/stage1/run_build.py --builder-tag --dry-run # prints the plan, no key needed PYTHONPATH=src python3 src/stage1/run_build.py --builder-tag --out-dir results/ # then score results//agent as above ``` ## Scenario fields (`scenarios`) | Field | Type | Description | |---|---|---| | `scenario_id` | string | Stable identifier, equal to the scenario file name | | `family` | string | Framework family: `permission`, `file`, `state`, `direct`, `headless`, `tui`, `artifact`, `startup`, `extension` | | `category` | string | Category name used in the paper (e.g. `permissions`, `session state`, `interactive terminal`) | | `role` | string | `test_case` (scored) or `positive_control` (not scored) | | `scored` | bool | Whether the scenario counts toward the pass rate | | `difficulty` | string | `standard` or `hard` | | `gated_by` | list[string] | Positive controls that must PASS for this test case's status to count | | `requires` | list[string] | Other scenarios this one depends on | | `summary` | string | One-line purpose of the scenario | | `description` | string | Purpose and step-by-step flow, from the scenario file | | `source_path` | string | Path of the scenario file in the GitHub repository | | `source_code` | string | Full source of the scenario (setup, stimulus, observation, and judgement) | A **scenario** is the session through which the mock model drives the candidate harness; a **test case** is a scenario plus its assertion. A **positive control** proves that the probed operation *can* succeed when nothing blocks it, so a blocked run in the gated test case reads as "the rule held" rather than "the command never worked". If you select a subset, keep it closed under `gated_by`. | Family | Total | Scored | Controls | |---|--:|--:|--:| | `file` | 17 | 9 | 8 | | `permission` | 16 | 8 | 8 | | `state` | 14 | 7 | 7 | | `headless` | 13 | 6 | 7 | | `direct` | 11 | 6 | 5 | | `tui` | 9 | 5 | 4 | | `artifact` | 7 | 7 | 0 | | `startup` | 5 | 4 | 1 | | `extension` | 2 | 1 | 1 | | **Total** | **94** | **53** | **41** | ## Leaderboard (`leaderboard`)

Pass rate of nineteen models

Pass rate (%) of 19 frontier models on the **full 328-case suite**. Each model built three independent candidate harnesses; `pass_rate` is their mean and `pass_rate_std` the standard deviation. Category columns are the per-category pass rates; `other` merges artifact checks, startup, and extension discovery. Positive controls are excluded from every denominator. GPT-6 Astra leads at 53.05%, the weakest model reaches 27.44%, and the cross-model mean is 38.30%. ## Citation ```bibtex @article{ge2026doc2harness, title = {Doc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents}, author = {Ge, Yuyao and Liu, Shenghua and Gao, Peizhong and Ge, Chendi and Liu, Chenxu and Wang, Yiwei and Mei, Lingrui and Bi, Baolong and Yao, Jiayu and Guo, Jiafeng and Cheng, Xueqi}, journal = {arXiv preprint arXiv:2610.20344}, year = {2026} } ``` ## License Apache 2.0, same as the [code repository](https://github.com/YuyaoGe/doc2harness). The contract skeleton derives from the project layout of [mini-swe-agent](https://github.com/SWE-agent/mini-swe-agent) (MIT); see `NOTICE` for upstream attribution.