Model-as-a-Harness; Harness Intelligence Model</p>\n<p>Website: <a href=\"https://bingreeky.github.io/JIT-site/\" rel=\"nofollow\">https://bingreeky.github.io/JIT-site/</a><br>GitHub: <a href=\"https://github.com/bingreeky/JIT\" rel=\"nofollow\">https://github.com/bingreeky/JIT</a><br>Hugging Face: <a href=\"https://huggingface.co/datasets/JIT-Agent/jit-meta-harness\">https://huggingface.co/datasets/JIT-Agent/jit-meta-harness</a></p>\n","updatedAt":"2026-08-27T03:07:30.737Z","author":{"_id":"6363a1fa123a5d5cd4a800e2","avatarUrl":"/avatars/a0961ca5463aae05de0b1574c0064fae.svg","fullname":"gbz","name":"greeky","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":5,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6455574035644531},"editors":["greeky"],"editorAvatarUrls":["/avatars/a0961ca5463aae05de0b1574c0064fae.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.25593","authors":[{"_id":"6a8fa9932c24e8c5fab32983","name":"Guibin Zhang","hidden":false},{"_id":"6a8fa9932c24e8c5fab32984","name":"Leo Lu","hidden":false},{"_id":"6a8fa9932c24e8c5fab32985","name":"Fangzhou Xie","hidden":false},{"_id":"6a8fa9932c24e8c5fab32986","name":"Kang Zhu","hidden":false},{"_id":"6a8fa9932c24e8c5fab32987","name":"Junhao Wang","hidden":false},{"_id":"6a8fa9932c24e8c5fab32988","name":"Zhifei Xie","hidden":false},{"_id":"6a8fa9932c24e8c5fab32989","name":"Zhaochen Yu","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298a","name":"Zihang Liu","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298b","name":"Zhongxiang Sun","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298c","name":"Qiankun Li","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298d","name":"Yue Liao","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298e","user":{"_id":"64c3698f007b906a75a14943","avatarUrl":"/avatars/da31d842225cc6074ecd904a82caf931.svg","isPro":false,"fullname":"Heng Chang","user":"hchang95","type":"user","name":"hchang95"},"name":"Heng Chang","status":"claimed_verified","statusLastChangedAt":"2026-08-27T08:45:04.952Z","hidden":false},{"_id":"6a8fa9932c24e8c5fab3298f","name":"Xiaobin Hu","hidden":false},{"_id":"6a8fa9932c24e8c5fab32990","name":"Qibing Ren","hidden":false},{"_id":"6a8fa9932c24e8c5fab32991","name":"Wangchunshu Zhou","hidden":false},{"_id":"6a8fa9932c24e8c5fab32992","name":"Shuicheng Yan","hidden":false}],"publishedAt":"2026-08-26T00:00:00.000Z","submittedOnDailyAt":"2026-08-27T00:00:00.000Z","title":"JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolution","submittedOnDailyBy":{"_id":"6363a1fa123a5d5cd4a800e2","avatarUrl":"/avatars/a0961ca5463aae05de0b1574c0064fae.svg","isPro":false,"fullname":"gbz","user":"greeky","type":"user","name":"greeky"},"summary":"Agent capability is not determined by the model alone. The agent harness, encompassing memory management, planning strategy, action protocol, and tool/skill orchestration, can dominate the contribution of the underlying foundation model. Yet harness design remains manual, task-specific, and fundamentally unscalable. We present JIT-Agent, a harness intelligence model trained to synthesize task-adaptive agent harnesses on the fly for arbitrary off-the-shelf agentic LLMs. We formalize the agent harness as a composable, machine-generatable artifact governed by a fixed four-module protocol, and train JIT-Agent to customize harnesses for a given task at hand, repair harnesses for stable and reliable execution, and self-evolve by distilling performance signals from an expanding archive of prior harness configurations. Equipped with JIT-Agent as a harness helper, DeepSeek-V4-Flash surpasses GPT-5.6 on DeepSearchQA (+9.1) and OdysseyBench (+4.3), while the already strong GLM-5.2 gains up to +20.2 points. Across controlled evaluations, JIT-Agent-generated harnesses are performance-competitive with mature agent runtimes such as OpenCode and Claude Code and consistently improve multi-scale model families of DeepSeek V4, Mimo-V2.5, and Qwen3.6. To our knowledge, JIT-Agent is the first model purpose-built for just-in-time harness generation, establishing harness intelligence as a trainable, transferable, and compounding dimension of agent capability orthogonal to model scaling.","upvotes":38,"discussionId":"6a8fa9932c24e8c5fab32993","projectPage":"https://bingreeky.github.io/JIT-site/","githubRepo":"https://github.com/bingreeky/JIT","githubRepoAddedBy":"user","ai_summary":"JIT-Agent is a trainable model that synthesizes adaptive agent harnesses for off-the-shelf LLMs, improving performance across diverse models and tasks.","ai_keywords":["JIT-Agent","agent harness","harness intelligence","composable harness","four-module protocol","self-evolve","distillation","DeepSearchQA","OdysseyBench","OpenCode","Claude Code"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":0,"organization":{"_id":"6508ab2b349930913196378b","name":"NationalUniversityofSingapore","fullname":"National University of Singapore","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/630ca0817dacb93b33506ce7/ZYUmpSMsa5Whihw3me2Bw.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6363a1fa123a5d5cd4a800e2","avatarUrl":"/avatars/a0961ca5463aae05de0b1574c0064fae.svg","isPro":false,"fullname":"gbz","user":"greeky","type":"user"},{"_id":"65082baabc8788c4064d5360","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/NZpFVnTpPGcCe8mvFMD-L.jpeg","isPro":false,"fullname":"Xiangyuan Xue","user":"xxyQwQ","type":"user"},{"_id":"685d283d22cb18bd944fd656","avatarUrl":"/avatars/fd97ea1042446688c9702a491f625fe4.svg","isPro":false,"fullname":"Nanfu Liu","user":"liunanfu1992","type":"user"},{"_id":"686e506e573b80f69c3bd7d7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/95CEjV6TRp3ah7gnSF9_5.png","isPro":false,"fullname":"EdwinMars","user":"EdwinYue","type":"user"},{"_id":"67c7d44419b236e0565358f4","avatarUrl":"/avatars/36f6dfd8c8dcdd69d8a9af5e58d978a4.svg","isPro":false,"fullname":"Zihang Liu","user":"zh-liu799","type":"user"},{"_id":"66559fa3a2d7a882a89167c0","avatarUrl":"/avatars/7d9d741da8baf054f205d8d0dafbd05f.svg","isPro":false,"fullname":"Yifan Hu","user":"hhh000","type":"user"},{"_id":"6309bfdab8d7b3889319b588","avatarUrl":"/avatars/572acdad470f765ef2e058ead3741e24.svg","isPro":false,"fullname":"SunZX","user":"Jeryi","type":"user"},{"_id":"66ea643899af9ac3463639b1","avatarUrl":"/avatars/252d470e761a57834dee3dbc60dfefed.svg","isPro":false,"fullname":"Disen Lan","user":"landisen","type":"user"},{"_id":"628c8598ef14f971b698107f","avatarUrl":"/avatars/3a4ad87e6b5f9e836a1160d869df1447.svg","isPro":false,"fullname":"Zhou","user":"Wangchunshu","type":"user"},{"_id":"66c993c9315af068a9f2bdc1","avatarUrl":"/avatars/34389705fa7c32d96e6166f0678e303a.svg","isPro":false,"fullname":"王文杰","user":"JackWang123","type":"user"},{"_id":"67bde7dec73e0b462c34d379","avatarUrl":"/avatars/f7656adc28805490124b6ed73fe73858.svg","isPro":false,"fullname":"Muxin Fu","user":"KANABOON1","type":"user"},{"_id":"6717c5c36bc2876059ed23ab","avatarUrl":"/avatars/52c68fb315760df5ef9323cd8ada5a3c.svg","isPro":false,"fullname":"Xin Zhou","user":"LMD0311","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6508ab2b349930913196378b","name":"NationalUniversityofSingapore","fullname":"National University of Singapore","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/630ca0817dacb93b33506ce7/ZYUmpSMsa5Whihw3me2Bw.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.25593.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2608.25593 in a dataset README.md to link it from this page.

Cite arxiv.org/abs/2608.25593 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.