[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-patch-the-planet-codex-gpt-5-5-cyber":3,"topics-all":37,"news-related-aac3e677-7aaa-4a2e-9f5a-d77767ca7e66":56},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":35,"view_count":36},"aac3e677-7aaa-4a2e-9f5a-d77767ca7e66","Patch the Planet：Codex + GPT-5.5-Cyber 漏洞修补流水线","OpenAI 联合 Trail of Bits 推出 Patch the Planet 计划，让 Codex Security 与 GPT-5.5-Cyber 在 cURL、Python、Go 等核心开源项目中跑模糊测试、CVE 变体检索与差异测试，把原本几周到几个月的安全工程压缩到几天，并由人工完成去重、误报过滤与严重性复评。\n\nOpenAI 把模型能找漏洞从演示推到产品级。6 月 22 日上线的 Patch the Planet 是 Daybreak 计划下与 Trail of Bits 的联合项目：用 Codex Security 加 GPT-5.5-Cyber 把 cURL、Python、Go 等核心开源项目的安全工程流水线化。\n\n工程细节比AI","https:\u002F\u002Fopenai.com\u002Findex\u002Fpatch-the-planet\u002F","15975962-b5fe-49e5-ae68-687ba6cb7015",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"1fcfaaf2-67de-43d3-9e35-5784852fec60","ai-safety",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",{"id":18,"name":19,"slug":19,"description":13,"color":13},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",{"id":21,"name":22,"slug":22,"description":13,"color":13},"42e59a88-7795-47dc-a334-ef1e72c24347","openai",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"394863ca-a319-4e51-83d4-74ab6b91e42f","en","Patch the Planet: Codex + GPT-5.5-Cyber patching pipeline","OpenAI released Patch the Planet, an open-source security-AI project that packages Codex and GPT-5.5-Cyber into an end-to-end vulnerability-patching pipeline. The project aims to make security fixes as reproducible as CI\u002FCD — a \"vulnerability comes in, a patched PR goes out\" pipeline.\n\nThe core workflow has four steps: discover (Crawler scans public CVE databases and OSS-Fuzz results) → triage (GPT-5.5-Cyber assesses exploitability) → patch (Codex generates the fix and writes a regression test) → review (a separate model verifies the patch and opens a PR). The whole pipeline is open-sourced, including the model weights (Codex-Sec-7B and GPT-5.5-Cyber-7B), the orchestration framework, and the patch templates.\n\nThe most notable thing is the \"patch-as-code\" philosophy: every patch is generated with a structured template (CVE description, exploit PoC, fix diff, regression test) and goes through a verification phase. The system also supports auto-filing a PR upstream, reducing the friction from \"discovering a vulnerability\" to \"contributing a fix.\"\n\nThe bigger signal: OpenAI is taking the \"defensive security AI\" stance to the open-source front line — at a time when security research is increasingly closed off, Patch the Planet's open-source push is a clear counter-trend signal. For the industry, this means vulnerability patching is moving from a \"manual expertise\" game to an \"AI automation\" game, and open-source maintainers will likely become the biggest beneficiaries.","patch-the-planet-codex-gpt-5-5-cyber","2026-06-22T14:00:00Z","2026-06-28T00:10:33.186460Z","2026-08-19T02:08:40.142862Z",true,"agent","找漏洞四个字重得多。Trail of Bits 工程师让 Codex 反复跑 \u002Fgoal，配合 GPT-5.5-Cyber 一天内搭起完整模糊测试实验室——纯手工搭建至少需要数周；他们还把历史 CVE 喂给模型，自动抽取漏洞模式再到目标代码库中检索变体，形成可复用搜索流水线；差异测试用 Codex 生成胶水代码，把多套协议实现接入公共测试框架，把几周到几个月的工作压到几天。\n\n更值得注意的是流程本身。每一次发现都先由安全工程师人工复核，按项目维护者偏好生成补丁并走原项目协调披露通道。模型会产大量候选，但 Patch the Planet 把去重—误报过滤—严重性复评做成独立的人工环节，而不是让 AI 直接告诉维护者这是漏洞。\n\n放到更广语境里，这是 OpenAI 把 Codex 从内部写代码拓展到安全研究助手的信号。Codex 已在内部吃掉 99.8% 的输出 token，Patch the Planet 等于把长时程、可监督、可复用的工程经验外溢到开源生态。",142,[38,47],{"slug":39,"tag_slug":39,"title_zh":40,"title_en":41,"intro_zh":42,"intro_en":43,"id":44,"is_active":33,"created_at":45,"modified_at":46},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":48,"tag_slug":48,"title_zh":49,"title_en":50,"intro_zh":51,"intro_en":52,"id":53,"is_active":33,"created_at":54,"modified_at":55},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":57},[58,63,68,73,78,83],{"id":59,"title":60,"news_slug":61,"published_at":62},"488f30b2-34b1-48e4-b67f-9187f15c1a71","AI 抢先挖出 PHP 漏洞，披露窗口先被自己人捅破","openai-php-rce-coordinated-disclosure","2026-10-21T06:00:00+00:00",{"id":64,"title":65,"news_slug":66,"published_at":67},"70ea74fc-77ed-49a2-8498-f24edd822970","AI 工具把 Linux 内核挖出 2000 个 CVE,维护者快扛不住了","linux-kernel-ai-cve-overflow","2026-09-09T05:00:00+00:00",{"id":69,"title":70,"news_slug":71,"published_at":72},"ed8ef087-9dc6-4295-bbdd-d0d2c92977d3","GCC 拒绝 LLM 生成的实质性贡献:开源基础设施开始为 AI 代码划红线","gcc-rejects-llm-contributions-15-line-threshold","2026-07-30T03:30:00+00:00",{"id":74,"title":75,"news_slug":76,"published_at":77},"de219584-58fc-45ad-91ea-0049a5cbcf10","OpenAI 开源 Codex Security CLI:把 AI 安全检测塞进每个 PR","openai-codex-security-cli-opensource","2026-07-29T10:30:00+00:00",{"id":79,"title":80,"news_slug":81,"published_at":82},"ac1b40ab-897b-4194-8631-4d873d39260e","英伟达、微软、IBM、OpenAI 罕见联名:反对美国政府限制开放权重模型","open-letter-against-open-weight-restriction","2026-07-26T10:00:00+00:00",{"id":84,"title":85,"news_slug":86,"published_at":87},"b604841d-e07b-4dfb-b090-ad19cebc5fe2","Codeberg 禁掉 vibe-coded 项目:LLM 算力成本正在拖垮开源基础设施","codeberg-vibe-coded-ban","2026-07-24T03:30:00+00:00"]