Repo0 doesn’t just generate code—it builds the architecture as it goes. It continuously reshapes the repository with cohesion and coupling until the structure truly makes sense.</p>\n","updatedAt":"2026-08-21T02:49:58.746Z","author":{"_id":"65684c80a9a1a6a50d779f58","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65684c80a9a1a6a50d779f58/it534ZdH5LxRub1M_o3uM.jpeg","fullname":"Silin Chen","name":"Silin-Chen","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.944898784160614},"editors":["Silin-Chen"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/65684c80a9a1a6a50d779f58/it534ZdH5LxRub1M_o3uM.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.19854","authors":[{"_id":"6a87bc7c89e517cbfd75dc0d","name":"Silin Chen","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc0e","name":"Haoyi Teng","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc0f","name":"Xiaodong Gu","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc10","name":"Yuling Shi","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc11","name":"Jiale Huang","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc12","name":"Yongpan Wang","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc13","name":"Hongyu Zhang","hidden":false},{"_id":"6a87bc7c89e517cbfd75dc14","name":"Haibing Guan","hidden":false}],"publishedAt":"2026-08-20T00:00:00.000Z","submittedOnDailyAt":"2026-08-21T00:00:00.000Z","title":"Repo0: Design-Driven Zero-to-All Code Generation","submittedOnDailyBy":{"_id":"65684c80a9a1a6a50d779f58","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65684c80a9a1a6a50d779f58/it534ZdH5LxRub1M_o3uM.jpeg","isPro":false,"fullname":"Silin Chen","user":"Silin-Chen","type":"user","name":"Silin-Chen"},"summary":"Large language model agents have made substantial progress in code generation, yet most existing systems assume a predefined repository architecture. This assumption does not hold in zero-to-all code generation, where an agent must construct an entire software project directly from natural-language requirements while maintaining a modular repository architecture throughout development. We present Repo0, a continuous structural evolution framework for zero-to-all code generation. Repo0 maintains an explicit architectural state instantiated as a Dual-Directed-Acyclic-Graph (Dual-DAG), consisting of a requirement-level DAG, a component-level DAG, and their alignment relation. Starting from natural-language requirements, it iteratively evolves component boundaries through structural actions guided by modularity metrics until structural convergence, after which the converged architecture guides test-driven development code generation. We evaluate Repo0 on six real-world repositories from RepoCraft using GPT-5 mini and DeepSeek V3.2. Repo0 achieves the highest Functionality Coverage and Pass Rate across all settings. Compared with RPG, the strongest repository-planning baseline, Repo0 improves Functionality Coverage by up to 20.08 percentage points and Pass Rate by up to 29.74 percentage points. Ablation and structural-evolution analyses further demonstrate the importance of the Dual-DAG architectural state, modularity-guided structural evolution, and explicit structural convergence.","upvotes":5,"discussionId":"6a87bc7d89e517cbfd75dc15","ai_summary":"Repo0 uses a dual-graph architectural state and modularity-guided structural evolution to generate complete software repositories from natural-language requirements with high functionality coverage.","ai_keywords":["zero-to-all code generation","Dual-DAG","requirement-level DAG","component-level DAG","structural actions","modularity metrics","structural convergence","test-driven development","RepoCraft"],"ai_summary_model":"thinkingmachines/Inkling-Small","organization":{"_id":"63e5ef7bf2e9a8f22c515654","name":"SJTU","fullname":"Shanghai Jiao Tong University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1676013394657-63e5ee22b6a40bf941da0928.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65684c80a9a1a6a50d779f58","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65684c80a9a1a6a50d779f58/it534ZdH5LxRub1M_o3uM.jpeg","isPro":false,"fullname":"Silin Chen","user":"Silin-Chen","type":"user"},{"_id":"69f56b3c1f33b9e6cfb990af","avatarUrl":"/avatars/41288ccbb7a24be49768fe5b05745710.svg","isPro":false,"fullname":"Adam Huang","user":"unicornshjl","type":"user"},{"_id":"67062b238587c775e4f14191","avatarUrl":"/avatars/e6cc278cff345048b195e62df7b99647.svg","isPro":false,"fullname":"Lihan","user":"SJTU-furrina","type":"user"},{"_id":"63ac5701c21e60a3e9b58aa7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ac5701c21e60a3e9b58aa7/g6EX7diOpuA94R2ab-rZC.png","isPro":true,"fullname":"Dipankar Sarkar","user":"dipankarsarkar","type":"user"},{"_id":"66d8512c54209e9101811e8e","avatarUrl":"/avatars/62dfd8e6261108f2508efe678d5a2a57.svg","isPro":false,"fullname":"M Saad Salman","user":"MSS444","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"63e5ef7bf2e9a8f22c515654","name":"SJTU","fullname":"Shanghai Jiao Tong University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1676013394657-63e5ee22b6a40bf941da0928.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.19854.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2608.19854 in a model README.md to link it from this page.

Cite arxiv.org/abs/2608.19854 in a dataset README.md to link it from this page.

Cite arxiv.org/abs/2608.19854 in a Space README.md to link it from this page.

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