{"id":1035,"date":"2025-08-15T10:27:00","date_gmt":"2025-08-15T02:27:00","guid":{"rendered":"http:\/\/www.applicationsofllm.com\/?p=1035"},"modified":"2025-09-17T12:38:52","modified_gmt":"2025-09-17T04:38:52","slug":"%e5%a4%a7%e6%a8%a1%e5%9e%8b%e6%99%ba%e8%83%bd%e4%bd%93%e4%b8%8d%e6%ad%a2%e8%83%bd%e5%86%99%e4%bb%a3%e7%a0%81%ef%bc%8c%e8%bf%98%e8%83%bd%e8%a2%ab%e8%ae%ad%e7%bb%83%e6%88%90%e7%99%bd%e5%b8%bd%e9%bb%91","status":"publish","type":"post","link":"http:\/\/www.applicationsofllm.com\/index.php\/2025\/08\/15\/%e5%a4%a7%e6%a8%a1%e5%9e%8b%e6%99%ba%e8%83%bd%e4%bd%93%e4%b8%8d%e6%ad%a2%e8%83%bd%e5%86%99%e4%bb%a3%e7%a0%81%ef%bc%8c%e8%bf%98%e8%83%bd%e8%a2%ab%e8%ae%ad%e7%bb%83%e6%88%90%e7%99%bd%e5%b8%bd%e9%bb%91\/","title":{"rendered":"\u5927\u6a21\u578b\u667a\u80fd\u4f53\u4e0d\u6b62\u80fd\u5199\u4ee3\u7801\uff0c\u8fd8\u80fd\u88ab\u8bad\u7ec3\u6210\u767d\u5e3d\u9ed1\u5ba2"},"content":{"rendered":"\n

\u5f53\u4eba\u4eec\u8fd8\u5728\u60ca\u53f9\u5927\u6a21\u578b\u80fd\u5199\u4ee3\u7801\u3001\u80fd\u81ea\u52a8\u5316\u529e\u516c\u65f6\uff0c\u5b83\u4eec\u6b63\u5728\u6084\u7136\u8e0f\u5165\u4e00\u4e2a\u66f4\u654f\u611f\u3001\u66f4\u5371\u9669\u7684\u9886\u57df \u2014\u2014\u7f51\u7edc\u5b89\u5168<\/strong>\u3002<\/p>\n\n\n\n

\u60f3\u8c61\u4e00\u4e0b\uff0c\u5982\u679c AI \u4e0d\u53ea\u662f\u5199\u4ee3\u7801\u7684\u52a9\u624b\uff0c\u800c\u662f\u80fd\u591f\u50cf\u300c\u767d\u5e3d\u9ed1\u5ba2\u300d\u4e00\u6837\uff0c\u5728\u4e0d\u7834\u574f\u7cfb\u7edf\u7684\u524d\u63d0\u4e0b\u6a21\u62df\u653b\u51fb\u3001\u53d1\u73b0\u6f0f\u6d1e\u3001\u63d0\u51fa\u4fee\u590d\u5efa\u8bae\uff0c\u4f1a\u5e26\u6765\u600e\u6837\u7684\u6539\u53d8\uff1f<\/p>\n\n\n\n

\u8fd9\u4e2a\u95ee\u9898\uff0c\u6700\u8fd1\u7531 Amazon AWS AI \u7684 Q Developer \u56e2\u961f\u7ed9\u51fa\u4e86\u7b54\u6848\u3002\u4ed6\u4eec\u5728 arXiv \u4e0a\u540c\u65f6\u53d1\u5e03\u4e86\u4e24\u7bc7\u8bba\u6587\uff0c\u63d0\u51fa\u4e86\u8bad\u7ec3\u7f51\u7edc\u5b89\u5168\u5927\u6a21\u578b\u7684\u5168\u65b0\u65b9\u6cd5\uff1aCyber-Zero<\/strong> \u548c CTF-Dojo<\/strong>\u3002\u8fd9\u4e24\u9879\u7814\u7a76\u4e0d\u4ec5\u662f\u5b66\u672f\u63a2\u7d22\uff0c\u66f4\u50cf\u662f\u4e00\u6b21\u300c\u5b9e\u6218\u6f14\u7ec3\u300d\u7684\u9884\u544a\uff0c\u9884\u793a\u7740\u5927\u6a21\u578b\u667a\u80fd\u4f53\u6b63\u5728\u4ece\u901a\u7528\u4efb\u52a1\u8d70\u5411\u7f51\u7edc\u5b89\u5168\u7684\u524d\u7ebf\u3002<\/p>\n\n\n\n

\u8bba\u6587 1: Cyber-Zero: Training Cybersecurity Agents without Runtime<\/p>\n\n\n\n

\u94fe\u63a5: https:\/\/arxiv.org\/abs\/2508.00910<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u8bba\u6587 2: Training Language Model Agents to Find Vulnerabilities with CTF-Dojo<\/p>\n\n\n\n

\u94fe\u63a5\uff1ahttps:\/\/arxiv.org\/abs\/2508.18370<\/p>\n\n\n\n

\u7f51\u7edc\u5b89\u5168  \u5927\u6a21\u578b\u843d\u5730\u7684\u4e00\u5ea7\u7279\u6b8a\u5821\u5792<\/strong><\/p>\n\n\n\n

\u5728\u901a\u7528\u4efb\u52a1\u4e0a\uff0c\u5927\u6a21\u578b\u7684\u8bad\u7ec3\u5df2\u7ecf\u5f62\u6210\u4e86\u76f8\u5bf9\u6210\u719f\u7684\u8303\u5f0f\uff1a\u6d77\u91cf\u6570\u636e\u3001\u957f\u65f6\u95f4\u9884\u8bad\u7ec3\u3001\u518d\u7ecf\u8fc7\u5bf9\u9f50\u4e0e\u5fae\u8c03\u3002\u4f46\u7f51\u7edc\u5b89\u5168\u573a\u666f\u4e0d\u540c\uff0c\u5176\u6838\u5fc3\u96be\u70b9\u5728\u4e8e\u8bad\u7ec3\u73af\u5883\u4e0e\u6570\u636e\u7684\u9ad8\u5ea6\u654f\u611f\u6027\u3002<\/p>\n\n\n\n

\u4e8b\u5b9e\u4e0a\uff0c\u95ed\u6e90\u5927\u6a21\u578b\u5df2\u7ecf\u5728\u5b89\u5168\u653b\u9632\u65b9\u5411\u5c55\u73b0\u51fa\u4e00\u5b9a\u6f5c\u529b\u3002Google \u7684 Project Zero <\/strong>\u56e2\u961f\u5c31\u66fe\u4f7f\u7528 Gemini \u7cfb\u5217\u6a21\u578b\u63a2\u7d22\u6f0f\u6d1e\u53d1\u73b0\uff0c\u4e00\u4e9b\u521d\u521b\u516c\u53f8\u751a\u81f3\u5c1d\u8bd5\u6784\u5efa\u57fa\u4e8e\u95ed\u6e90\u6a21\u578b\u7684\u300cAI \u7ea2\u961f\u300d\uff0c\u7528\u6765\u6a21\u62df\u653b\u51fb\u5e76\u8fdb\u884c\u9632\u5fa1\u9a8c\u8bc1\u3002\u5b9e\u9645\u6848\u4f8b\u8868\u660e\uff0c\u8fd9\u4e9b\u5f3a\u5927\u7684\u95ed\u6e90\u6a21\u578b\u786e\u5b9e\u5177\u5907\u4e86\u53d1\u73b0\u6f0f\u6d1e\u3001\u81ea\u52a8\u5316\u6267\u884c\u653b\u51fb\u6b65\u9aa4\u7684\u6f5c\u529b\u3002<\/p>\n\n\n\n

\u7136\u800c\u95ee\u9898\u5728\u4e8e\uff0c\u8fd9\u4e9b\u6a21\u578b\u7684\u8bad\u7ec3\u8303\u5f0f\u548c\u6570\u636e\u96c6\u5b8c\u5168\u4e0d\u900f\u660e\u3002\u6211\u4eec\u65e0\u6cd5\u5f97\u77e5\u5b83\u4eec\u662f\u5982\u4f55\u4e60\u5f97\u653b\u9632\u77e5\u8bc6\u7684\uff0c\u4e5f\u65e0\u6cd5\u9a8c\u8bc1\u6a21\u578b\u7684\u5b89\u5168\u6027\u4e0e\u53ef\u9760\u6027\u3002\u66f4\u91cd\u8981\u7684\u662f\uff0c\u95ed\u6e90\u6a21\u578b\u65e0\u6cd5\u88ab\u7814\u7a76\u8005\u548c\u4f01\u4e1a\u5b89\u5168\u56e2\u961f\u81ea\u4e3b\u6539\u9020\u6216\u63a7\u5236\uff0c\u8fd9\u672c\u8eab\u5728\u5b89\u5168\u9886\u57df\u662f\u4e00\u79cd\u6f5c\u5728\u98ce\u9669\u3002<\/p>\n\n\n\n

\u53e6\u4e00\u65b9\u9762\uff0c\u5982\u679c\u8981\u8ba9\u6a21\u578b\u4ece\u96f6\u5f00\u59cb\u5b66\u4f1a\u653b\u9632\uff0c\u4f20\u7edf\u601d\u8def\u9700\u8981\u642d\u5efa\u771f\u5b9e\u8fd0\u884c\u73af\u5883\uff0c\u4ee5\u751f\u6210\u4ea4\u4e92\u8f68\u8ff9\u3002\u4f46\u8fd9\u79cd\u65b9\u5f0f\u6210\u672c\u9ad8\u3001\u98ce\u9669\u5927\uff0c\u751a\u81f3\u53ef\u80fd\u5728\u5b9e\u9a8c\u4e2d\u89e6\u53d1\u4e0d\u53ef\u63a7\u7684\u653b\u51fb\u3002\u800c\u9ad8\u8d28\u91cf\u7684\u5b89\u5168\u653b\u9632\u6570\u636e\u672c\u5c31\u6781\u5ea6\u7a00\u7f3a\u3002\u6f0f\u6d1e\u5229\u7528\u548c\u4fee\u590d\u5f80\u5f80\u6d89\u53ca\u590d\u6742\u7684\u73af\u5883\u72b6\u6001\u3001\u7cfb\u7edf\u8c03\u7528\u548c\u957f\u65f6\u95f4\u63a8\u7406\uff0c\u5f88\u96be\u50cf\u81ea\u7136\u8bed\u8a00\u6587\u672c\u90a3\u6837\u5bb9\u6613\u8f6c\u5316\u4e3a\u6807\u51c6\u8bed\u6599\u3002<\/p>\n\n\n\n

\u8fd9\u610f\u5473\u7740\uff0c\u5982\u679c\u7ee7\u7eed\u6cbf\u7528\u4f20\u7edf\u65b9\u5f0f\uff0c\u300cAI \u767d\u5e3d\u9ed1\u5ba2\u300d\u53ef\u80fd\u6c38\u8fdc\u53ea\u80fd\u505c\u7559\u5728\u5b9e\u9a8c\u9636\u6bb5\u3002Amazon \u56e2\u961f\u6b63\u662f\u7784\u51c6\u4e86\u8fd9\u4e2a\u74f6\u9888\uff0c\u63d0\u51fa\u4e86\u4e24\u4e2a\u4e92\u8865\u7684\u89e3\u51b3\u65b9\u6848\uff1aCyber-Zero \u81f4\u529b\u4e8e\u300c\u5982\u4f55\u751f\u6210\u5b89\u5168\u800c\u9ad8\u6548\u7684\u8bad\u7ec3\u6570\u636e\u300d\uff0c\u800c CTF-Dojo \u5219\u4e13\u6ce8\u4e8e\u300c\u5982\u4f55\u5728\u5b9e\u6218\u4e2d\u8bad\u7ec3\u6a21\u578b\u53d1\u73b0\u6f0f\u6d1e\u300d\u3002<\/p>\n\n\n\n

Cyber-Zero  \u65e0\u9700\u771f\u5b9e\u73af\u5883\u7684\u6a21\u62df\u8bad\u7ec3\u573a<\/strong><\/p>\n\n\n\n

Cyber-Zero \u7684\u6838\u5fc3\u601d\u60f3\u662f\u300cruntime-free training<\/strong>\u300d\uff0c\u5373\u5b8c\u5168\u4e0d\u4f9d\u8d56\u771f\u5b9e\u8fd0\u884c\u73af\u5883\uff0c\u800c\u662f\u901a\u8fc7\u5df2\u6709\u77e5\u8bc6\u548c\u8bed\u8a00\u5efa\u6a21\u751f\u6210\u8bad\u7ec3\u6240\u9700\u7684\u9ad8\u8d28\u91cf\u884c\u4e3a\u8f68\u8ff9 (trajectories)\u3002<\/p>\n\n\n\n

\u56e2\u961f\u6ce8\u610f\u5230\uff0c\u516c\u5f00\u7684 CTF\uff08Capture The Flag\uff09\u7ade\u8d5b writeups \u662f\u6781\u5176\u5b9d\u8d35\u7684\u8d44\u6e90\u3002\u5b83\u4eec\u8bb0\u5f55\u4e86\u53c2\u8d5b\u8005\u5982\u4f55\u5206\u6790\u9898\u76ee\u3001\u5c1d\u8bd5\u653b\u51fb\u3001\u5b9a\u4f4d\u6f0f\u6d1e\u4ee5\u53ca\u6700\u7ec8\u89e3\u9898\u7684\u8fc7\u7a0b\u3002Cyber-Zero \u6b63\u662f\u57fa\u4e8e\u8fd9\u4e9b writeups\uff0c\u6784\u5efa\u51fa\u9ad8\u8d28\u91cf\u7684\u8bad\u7ec3\u8f68\u8ff9\u3002<\/p>\n\n\n\n

\u5177\u4f53\u6765\u8bf4\uff0c\u7cfb\u7edf\u9996\u5148\u4ece writeups \u4e2d\u63d0\u53d6\u5173\u952e\u6b65\u9aa4\u548c\u601d\u8def\uff0c\u7136\u540e\u901a\u8fc7\u8bbe\u5b9a\u4e0d\u540c\u7684\u4eba\u683c\uff08persona\uff09\uff0c\u8ba9\u5927\u6a21\u578b\u5728\u7eaf\u6587\u672c\u73af\u5883\u4e2d\u6a21\u62df\u653b\u9632\u53cc\u65b9\u7684\u5bf9\u8bdd\u4e0e\u64cd\u4f5c\u3002\u4f8b\u5982\uff0c\u653b\u51fb\u8005 persona \u4f1a\u751f\u6210\u53ef\u80fd\u7684\u5229\u7528\u8def\u5f84\uff0c\u9632\u5fa1\u8005 persona \u4f1a\u8fdb\u884c\u5e94\u5bf9\u3002\u8fd9\u4e00\u8fc7\u7a0b\u4e2d\u751f\u6210\u7684\u957f\u5e8f\u5217\u4ea4\u4e92\u88ab\u89c6\u4f5c\u884c\u4e3a\u8f68\u8ff9\uff0c\u7528\u4e8e\u8bad\u7ec3\u7f51\u7edc\u5b89\u5168\u667a\u80fd\u4f53\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u5b9e\u9a8c\u8868\u660e\uff0c\u8fd9\u79cd\u514d\u8fd0\u884c\u65f6\u7684\u8f68\u8ff9\u751f\u6210\u4e0d\u4ec5\u89c4\u6a21\u53ef\u89c2\uff0c\u800c\u4e14\u591a\u6837\u6027\u4e30\u5bcc\uff0c\u8986\u76d6\u4e86\u5e38\u89c1\u7684\u653b\u9632\u6a21\u5f0f\u3002\u4e0e\u771f\u5b9e\u73af\u5883\u751f\u6210\u7684\u8f68\u8ff9\u76f8\u6bd4\uff0cCyber-Zero \u7684\u6570\u636e\u5728\u6f0f\u6d1e\u5b9a\u4f4d\u3001\u653b\u51fb\u8def\u5f84\u63a8\u7406\u7b49\u4efb\u52a1\u4e0a\u7684\u8bad\u7ec3\u6548\u679c\u6beb\u4e0d\u900a\u8272\uff0c\u751a\u81f3\u5728\u90e8\u5206\u6307\u6807\u4e0a\u8868\u73b0\u66f4\u4f18\u3002\u8fd9\u610f\u5473\u7740\uff0cAI \u767d\u5e3d\u9ed1\u5ba2\u53ef\u4ee5\u5728\u4e00\u4e2a\u5b8c\u5168\u5b89\u5168\u7684\u865a\u62df\u8bad\u7ec3\u8425\u4e2d\u53cd\u590d\u4f18\u5316\uff0c\u800c\u4e0d\u5fc5\u62c5\u5fc3\u6210\u672c\u548c\u98ce\u9669\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u56e2\u961f\u8fd8\u5f97\u51fa\u51e0\u9879\u5173\u952e\u53d1\u73b0\uff1a<\/p>\n\n\n\n

    \n
  • \u901a\u7528\u7684\u8f6f\u4ef6\u5de5\u7a0b\u667a\u80fd\u4f53\uff08SWE Agents\uff09\u65e0\u6cd5\u76f4\u63a5\u8fc1\u79fb\u81f3\u7f51\u7edc\u5b89\u5168\u4efb\u52a1\u3002\u5199\u4ee3\u7801 \u2260 \u627e\u6f0f\u6d1e\uff0c\u4e24\u7c7b\u6280\u80fd\u4e4b\u95f4\u5b58\u5728\u660e\u663e\u9e3f\u6c9f\u3002<\/li>\n\n\n\n
  • \u6a21\u578b\u89c4\u6a21\u4e0e\u6027\u80fd\u5bc6\u5207\u76f8\u5173\uff1a\u53c2\u6570\u66f4\u5927\u7684\u6a21\u578b\u66f4\u64c5\u957f\u7ef4\u6301\u957f\u7a0b\u63a8\u7406\u94fe\uff0c\u8de8\u591a\u6b65\u7ec4\u5408\u547d\u4ee4\uff0c\u5e76\u5728\u591a\u8f6e\u4ea4\u4e92\u4e2d\u4fdd\u6301\u72b6\u6001\u8fde\u8d2f\uff0c\u8fd9\u5bf9\u590d\u6742\u653b\u9632\u81f3\u5173\u91cd\u8981\u3002<\/li>\n\n\n\n
  • \u7ecf\u8fc7 Cyber-Zero \u8f68\u8ff9\u5fae\u8c03\u7684 32B \u667a\u80fd\u4f53\uff0c\u6027\u80fd\u5df2\u63a5\u8fd1\u95ed\u6e90\u6a21\u578b Claude-3.7-Sonnet\uff0c\u800c\u63a8\u7406\u6210\u672c\u4ec5\u4e3a\u5176 1%\u3002<\/li>\n<\/ul>\n\n\n\n

    \u8fd9\u4e9b\u7ed3\u679c\u4e00\u65b9\u9762\u51f8\u663e\u4e86 Cyber-Zero \u7684\u5b9e\u7528\u4ef7\u503c\uff1a\u5b83\u4e0d\u4ec5\u80fd\u5b89\u5168\u3001\u4f4e\u6210\u672c\u5730\u751f\u6210\u8bad\u7ec3\u6570\u636e\uff0c\u8fd8\u80fd\u8ba9\u6a21\u578b\u901a\u8fc7\u5fae\u8c03\u5728\u5b89\u5168\u4efb\u52a1\u4e0a\u5177\u5907\u5b9e\u7528\u80fd\u529b\uff1b\u53e6\u4e00\u65b9\u9762\u4e5f\u6307\u51fa\u4e86\u7814\u7a76\u65b9\u5411\uff1a\u5982\u679c\u4e0d\u9488\u5bf9\u5b89\u5168\u4efb\u52a1\u8fdb\u884c\u4e13\u95e8\u4f18\u5316\uff0c\u5373\u4fbf\u662f\u5927\u89c4\u6a21\u7684\u901a\u7528 SWE \u667a\u80fd\u4f53\u4e5f\u96be\u4ee5\u627f\u62c5\u767d\u5e3d\u9ed1\u5ba2\u7684\u89d2\u8272\u3002<\/p>\n\n\n\n

    CTF-Dojo  \u8ba9 AI \u5728\u5b9e\u6218\u4e2d\u5b66\u4f1a\u53d1\u73b0\u6f0f\u6d1e<\/strong><\/p>\n\n\n\n

    \u5982\u679c\u8bf4 Cyber-Zero<\/strong> \u63d0\u4f9b\u7684\u662f\u4e00\u4e2a\u300c\u865a\u6784\u7684\u8bad\u7ec3\u573a\u300d\uff0c\u5b83\u901a\u8fc7\u89e3\u6790 CTF writeups \u4e0e persona \u6a21\u62df\uff0c\u5728\u7eaf\u6587\u672c\u7a7a\u95f4\u4e2d\u751f\u6210\u653b\u9632\u8f68\u8ff9\uff0c\u8ba9\u6a21\u578b\u5728\u5b8c\u5168\u65e0\u98ce\u9669\u7684\u865a\u62df\u73af\u5883\u4e2d\u5b66\u4e60\uff1b\u90a3\u4e48 CTF-Dojo<\/strong> \u5c31\u662f\u4e00\u4e2a\u300c\u771f\u5b9e\u7684\u6218\u573a\u300d\u3002\u5b83\u76f4\u63a5\u6784\u5efa\u53ef\u8fd0\u884c\u7684 CTF \u653b\u9632\u73af\u5883\uff0c\u8ba9\u667a\u80fd\u4f53\u80fd\u591f\u771f\u6b63\u6267\u884c\u547d\u4ee4\u3001\u4e0e\u7cfb\u7edf\u4ea4\u4e92\u3001\u53d1\u73b0\u5e76\u5229\u7528\u6f0f\u6d1e\u3002\u524d\u8005\u5f3a\u8c03\u89c4\u6a21\u5316\u3001\u5b89\u5168\u3001\u9ad8\u6548\u7684\u6570\u636e\u751f\u6210\uff0c\u540e\u8005\u5f3a\u8c03\u8d34\u8fd1\u5b9e\u6218\u7684\u653b\u9632\u6f14\u7ec3\uff0c\u4e24\u8005\u4e00\u865a\u4e00\u5b9e\uff0c\u5f62\u6210\u4e92\u8865\u3002<\/p>\n\n\n\n

    CTF-Dojo \u7684\u6838\u5fc3\u96be\u70b9\u5728\u4e8e\uff1a\u5982\u4f55\u5728\u5927\u89c4\u6a21\u4e0b\u4e3a LLM \u667a\u80fd\u4f53\u63d0\u4f9b\u7a33\u5b9a\u7684\u8fd0\u884c\u73af\u5883\u3002\u4f20\u7edf SWE\uff08\u8f6f\u4ef6\u5de5\u7a0b\uff09\u4ee3\u7406\u901a\u5e38\u9700\u8981\u4e13\u5bb6\u624b\u52a8\u914d\u7f6e\u73af\u5883\u624d\u80fd\u8fd0\u884c\uff0c\u800c\u6bcf\u4e2a\u4efb\u52a1\u7684\u51c6\u5907\u5de5\u4f5c\u5f80\u5f80\u8017\u65f6\u6570\u5468\uff0c\u6781\u5927\u9650\u5236\u4e86\u7814\u7a76\u89c4\u6a21\u3002\u4e3a\u6b64\uff0cAmazon \u56e2\u961f\u63d0\u51fa\u4e86 CTF-Forge<\/strong>\uff0c\u4e00\u79cd\u80fd\u591f\u5728\u51e0\u5206\u949f\u5185\u81ea\u52a8\u642d\u5efa\u8fd0\u884c\u65f6\u7684\u5bb9\u5668\u5316\u5de5\u5177\uff0c\u53ef\u4ee5\u5feb\u901f\u90e8\u7f72\u6570\u767e\u4e2a\u6311\u6218\u5b9e\u4f8b\uff0c\u663e\u8457\u964d\u4f4e\u4e86\u4eba\u529b\u6210\u672c\u3002<\/p>\n\n\n\n

    \"\"<\/figure>\n\n\n\n

    \u7814\u7a76\u56e2\u961f\u9009\u62e9\u4e86\u5168\u7403\u6700\u5177\u4ee3\u8868\u6027\u7684 pwn.college CTF Archive<\/strong> \u4f5c\u4e3a\u79cd\u5b50\u6570\u636e\u3002\u8be5\u6570\u636e\u96c6\u6536\u5f55\u4e86\u6570\u767e\u4e2a\u6765\u81ea\u9876\u7ea7\u8d5b\u4e8b\u7684\u9ad8\u8d28\u91cf\u9898\u76ee\uff0c\u6db5\u76d6\u516d\u5927\u7c7b\u522b\uff0c\u4ece Web \u6f0f\u6d1e\u3001\u4e8c\u8fdb\u5236\u5229\u7528\u5230\u5bc6\u7801\u5b66\u6311\u6218\u4e00\u5e94\u4ff1\u5168\u3002\u901a\u8fc7\u7cbe\u5fc3\u7b5b\u9009\uff0c\u5e76\u6392\u9664\u6d4b\u8bd5\u6570\u636e\u4e2d\u5df2\u5305\u542b\u7684\u9898\u76ee\uff0c\u6700\u7ec8\u5f62\u6210\u4e86 658 <\/strong>\u4e2a\u72ec\u7acb\u4efb\u52a1\u5b9e\u4f8b<\/strong>\uff0c\u4e3a\u667a\u80fd\u4f53\u8bad\u7ec3\u63d0\u4f9b\u4e86\u575a\u5b9e\u7684\u57fa\u7840\u3002<\/p>\n\n\n\n

    \"\"<\/figure>\n\n\n\n

    \u7136\u800c\uff0c\u6700\u521d\u5b9e\u9a8c\u8868\u660e\uff0c\u5f00\u6e90\u6a21\u578b\u5728\u8fd9\u4e9b\u590d\u6742\u4efb\u52a1\u4e0a\u7684\u6210\u529f\u7387\u6781\u4f4e\u3002\u5927\u90e8\u5206 OSS \u6a21\u578b\u53ea\u80fd\u5b8c\u6210\u5c11\u6570\u6311\u6218\uff0c\u751f\u6210\u7684\u8f68\u8ff9\u4e5f\u8d28\u91cf\u53c2\u5dee\u4e0d\u9f50\u3002\u4e3a\u4e86\u63d0\u9ad8\u53ef\u7528\u6837\u672c\u7684\u4ea7\u51fa\u7387\uff0c\u56e2\u961f\u5f15\u5165\u4e86\u4e09\u9879\u63a8\u7406\u9636\u6bb5\u589e\u5f3a\u6280\u672f\uff1a<\/p>\n\n\n\n

      \n
    • \u5c06\u516c\u5f00\u7684\u8d5b\u9898\u7b14\u8bb0\uff08writeups\uff09 \u4f5c\u4e3a\u63d0\u793a\uff0c\u5e2e\u52a9\u6a21\u578b\u66f4\u5feb\u9501\u5b9a\u89e3\u9898\u65b9\u5411\uff1b<\/li>\n\n\n\n
    • \u8fd0\u884c\u65f6\u589e\u5f3a<\/strong>\uff1a\u901a\u8fc7\u5728\u6267\u884c\u8fc7\u7a0b\u4e2d\u52a8\u6001\u4fee\u6539\u73af\u5883\u914d\u7f6e\u6216\u4efb\u52a1\u7ea6\u675f\uff0c\u628a\u8fc7\u4e8e\u590d\u6742\u7684\u6311\u6218\u300c\u964d\u7ef4\u300d\uff0c\u4ece\u800c\u63d0\u5347\u6a21\u578b\u5b8c\u6210\u4efb\u52a1\u7684\u6210\u529f\u7387\uff1b<\/li>\n\n\n\n
    • \u6559\u5e08\u6a21\u578b\u591a\u6837\u5316<\/strong>\uff1a\u4e0d\u4ec5\u4f9d\u8d56\u5355\u4e00\u6a21\u578b\u751f\u6210\u89e3\u9898\u8f68\u8ff9\uff0c\u800c\u662f\u540c\u65f6\u8c03\u7528\u591a\u79cd\u4e0d\u540c\u7c7b\u578b\u7684\u5927\u6a21\u578b\uff08\u5305\u62ec\u5f00\u6e90\u548c\u95ed\u6e90\uff09\uff0c\u8ba9\u5b83\u4eec\u5404\u81ea\u8d21\u732e\u6210\u529f\u6848\u4f8b\uff0c\u4ee5\u6b64\u83b7\u5f97\u66f4\u4e30\u5bcc\u3001\u66f4\u5177\u591a\u6837\u6027\u7684\u8bad\u7ec3\u6837\u672c\u3002<\/li>\n<\/ul>\n\n\n\n

      \u6700\u7ec8\uff0c\u56e2\u961f\u4e3b\u8981\u4f9d\u8d56 Qwen3-Code-480B<\/strong> \u548c DeepSeek-V3-0324<\/strong> \u4e24\u4e2a\u5f3a\u5927\u7684\u5f00\u6e90\u6a21\u578b\uff0c\u6536\u96c6\u5230\u6765\u81ea 274 \u4e2a\u6311\u6218\u7684 1000+ \u6210\u529f\u8f68\u8ff9<\/strong>\u3002\u5728\u53bb\u9664\u5197\u4f59\u3001\u9650\u5236\u6bcf\u4e2a\u4efb\u52a1\u5b9e\u4f8b\u7684\u6700\u5927\u6837\u672c\u6570\u540e\uff0c\u6700\u7ec8\u5f97\u5230\u4e86 486 <\/strong>\u6761\u9ad8\u8d28\u91cf\u3001\u7ecf\u8fc7\u8fd0\u884c\u9a8c\u8bc1\u7684\u8f68\u8ff9<\/strong>\u3002<\/p>\n\n\n\n

      \u57fa\u4e8e\u8fd9\u4e9b\u6570\u636e\uff0c\u7814\u7a76\u56e2\u961f\u5bf9 Qwen3 \u7cfb\u5217\u6a21\u578b\uff088B\u300114B \u548c 32B \u53c2\u6570\u89c4\u6a21\uff09\u8fdb\u884c\u4e86\u8bad\u7ec3\uff0c\u5e76\u5728\u591a\u4e2a\u7f51\u7edc\u5b89\u5168\u57fa\u51c6\u4efb\u52a1\u4e0a\u8bc4\u4f30\u4e86\u6548\u679c\u3002\u7ed3\u679c\u663e\u793a\uff0c\u7ecf\u8fc7 CTF-Dojo \u8bad\u7ec3\u7684\u6a21\u578b\uff0c\u5728 EnIGMA+ \u57fa\u51c6\uff08\u6e90\u81ea\u524d\u4f5c Cyber-Zero\uff09\u4e0a\u53d6\u5f97\u4e86\u6700\u9ad8 11.6% \u7684\u7edd\u5bf9\u63d0\u5347\uff0c\u4e0d\u4ec5\u8d85\u8fc7\u4e86\u5f00\u6e90\u57fa\u7ebf\u6a21\u578b\uff0c\u8fd8\u8868\u73b0\u51fa\u4e0e\u95ed\u6e90\u6a21\u578b\u63a5\u8fd1\u7684\u6c34\u5e73\u3002\u66f4\u91cd\u8981\u7684\u662f\uff0c\u968f\u7740\u8bad\u7ec3\u6837\u672c\u6570\u91cf\u7684\u589e\u52a0\uff0c\u6027\u80fd\u5448\u73b0\u51fa\u6e05\u6670\u7684\u53ef\u6269\u5c55\u6027\uff0c\u8bc1\u660e\u4e86\u5728\u771f\u5b9e\u73af\u5883\u8f68\u8ff9\u9a71\u52a8\u4e0b\uff0c\u5927\u6a21\u578b\u5728\u7f51\u7edc\u5b89\u5168\u4efb\u52a1\u4e0a\u7684\u6f5c\u529b\u53ef\u4ee5\u88ab\u7cfb\u7edf\u6027\u6fc0\u53d1\u3002<\/p>\n\n\n\n

      \"\"<\/figure>\n\n\n\n

      \u8fd9\u4e9b\u7ed3\u679c\u610f\u5473\u7740\uff0cCTF-Dojo \u4e0d\u4ec5\u89e3\u51b3\u4e86\u8fc7\u53bb\u300c\u73af\u5883\u96be\u4ee5\u5927\u89c4\u6a21\u914d\u7f6e\u300d\u7684\u5de5\u7a0b\u96be\u9898\uff0c\u8fd8\u9a8c\u8bc1\u4e86\u4e00\u4e2a\u6838\u5fc3\u79d1\u5b66\u95ee\u9898\uff1a\u7f51\u7edc\u5b89\u5168\u667a\u80fd\u4f53\u7684\u6027\u80fd\u80fd\u591f\u968f\u7740\u6267\u884c\u6570\u636e\u7684\u589e\u52a0\u800c\u6301\u7eed\u63d0\u5347<\/strong>\u3002\u5728\u5df2\u6709 SWE \u4ee3\u7406\u65e0\u6cd5\u6cdb\u5316\u7684\u60c5\u51b5\u4e0b\uff0cCTF-Dojo \u7ed9\u51fa\u4e86\u4e00\u6761\u6e05\u6670\u7684\u9053\u8def\uff1a\u901a\u8fc7\u89c4\u6a21\u5316\u3001\u81ea\u52a8\u5316\u7684\u8fd0\u884c\u73af\u5883\u6536\u96c6\u8f68\u8ff9\uff0c\u63a8\u52a8\u6a21\u578b\u9010\u6b65\u903c\u8fd1\u4eba\u7c7b\u767d\u5e3d\u9ed1\u5ba2\u7684\u5b9e\u6218\u6c34\u5e73\u3002<\/p>\n\n\n\n

      \u4ece\u865a\u62df\u5230\u5b9e\u6218\u7684\u7ec4\u5408\u62f3<\/strong><\/p>\n\n\n\n

      \u628a Cyber-Zero \u548c CTF-Dojo \u653e\u5728\u4e00\u8d77\u770b\uff0c\u5c31\u4f1a\u53d1\u73b0\u5b83\u4eec\u5f62\u6210\u4e86\u4e00\u4e2a\u95ed\u73af\u3002Cyber-Zero \u63d0\u4f9b\u7684\u662f\u5b89\u5168\u3001\u53ef\u6269\u5c55\u7684\u8bad\u7ec3\u6570\u636e\u6765\u6e90\uff0c\u76f8\u5f53\u4e8e\u4e00\u4e2a\u865a\u62df\u8bad\u7ec3\u8425\uff1b\u800c CTF-Dojo \u5219\u662f\u5b9e\u6218\u6b66\u9986\uff0c\u8ba9\u6a21\u578b\u5728\u771f\u5b9e\u6311\u6218\u4e2d\u4e0d\u65ad\u8fed\u4ee3\u3002\u524d\u8005\u89e3\u51b3\u4e86\u6570\u636e\u4e0e\u6210\u672c\u7684\u95ee\u9898\uff0c\u540e\u8005\u89e3\u51b3\u4e86\u80fd\u529b\u4e60\u5f97\u4e0e\u8fc1\u79fb\u7684\u95ee\u9898\u3002\u4e24\u8005\u7ed3\u5408\uff0c\u4e3a AI \u767d\u5e3d\u9ed1\u5ba2\u7684\u6210\u957f\u63d0\u4f9b\u4e86\u5b8c\u6574\u8def\u5f84\u3002<\/p>\n\n\n\n

      \u8fd9\u79cd\u8bbe\u8ba1\u601d\u8def\u7684\u610f\u4e49\u5728\u4e8e\uff0c\u5b83\u4e0d\u4ec5\u8ffd\u6c42\u7406\u8bba\u4e0a\u7684\u53ef\u884c\u6027\uff0c\u8fd8\u5f3a\u8c03\u5728\u751f\u4ea7\u73af\u5883\u4e2d\u771f\u6b63\u53ef\u90e8\u7f72\u3002\u6b63\u5982\u8bba\u6587\u4e2d\u5c55\u793a\u7684\u5b9e\u9a8c\u7ed3\u679c\uff0cCyber-Zero \u7684\u6570\u636e\u751f\u6210\u548c CTF-Dojo \u7684\u73af\u5883\u6784\u5efa\u90fd\u80fd\u89c4\u6a21\u5316\u8fd0\u884c\uff0c\u4e14\u80fd\u5728\u771f\u5b9e\u4efb\u52a1\u4e0a\u5e26\u6765\u53ef\u9a8c\u8bc1\u7684\u6027\u80fd\u63d0\u5347\u3002\u8fd9\u6807\u5fd7\u7740 AI \u5728\u7f51\u7edc\u5b89\u5168\u65b9\u5411\u6b63\u5728\u9010\u6b65\u8fdb\u5165\u5e94\u7528\u843d\u5730\u9636\u6bb5\u3002<\/p>\n\n\n\n

      \u672a\u6765\u610f\u4e49\u4e0e\u6311\u6218<\/strong><\/p>\n\n\n\n

      AI \u767d\u5e3d\u9ed1\u5ba2\u8574\u85cf\u5e7f\u9614\u524d\u666f\uff1a\u5728\u4f01\u4e1a\u5b89\u5168\u56e2\u961f\u4e2d\uff0c\u5b83\u53ef\u4ee5\u4f5c\u4e3a\u865a\u62df\u6210\u5458\uff0c\u81ea\u52a8\u626b\u63cf\u4ee3\u7801\u3001\u53d1\u73b0\u6f5c\u5728\u6f0f\u6d1e\uff0c\u5e76\u63d0\u51fa\u4fee\u590d\u5efa\u8bae\uff1b\u5728\u7ea2\u961f\u6f14\u7ec3\u4e2d\uff0c\u5b83\u53ef\u4ee5\u5145\u5f53\u5bf9\u624b\u89d2\u8272\uff0c\u5e2e\u52a9\u6d4b\u8bd5\u9632\u5fa1\u7cfb\u7edf\uff1b\u5728\u6559\u80b2\u573a\u666f\u4e2d\uff0c\u5b83\u53ef\u4ee5\u6210\u4e3a\u5b66\u5458\u7684\u300c\u966a\u7ec3\u300d\uff0c\u63d0\u4f9b\u4e2a\u6027\u5316\u7684\u6311\u6218\u548c\u53cd\u9988\u3002\u66f4\u957f\u8fdc\u6765\u770b\uff0c\u968f\u7740\u6210\u672c\u964d\u4f4e\u548c\u6280\u672f\u6210\u719f\uff0c\u4e2d\u5c0f\u4f01\u4e1a\u4e5f\u6709\u671b\u501f\u52a9\u8fd9\u6837\u7684\u7cfb\u7edf\u83b7\u5f97\u300c\u666e\u60e0\u5b89\u5168\u300d\u3002<\/p>\n\n\n\n

      \u4f46\u4e0e\u6b64\u540c\u65f6\uff0c\u8fd9\u9879\u6280\u672f\u7684\u53cc\u91cd\u7528\u9014\u5c5e\u6027<\/strong>\u4e0d\u5bb9\u5ffd\u89c6\u3002\u6b63\u5982\u7814\u7a76\u56e2\u961f\u5728\u8bba\u6587\u4e2d\u5f3a\u8c03\u7684\u90a3\u6837\uff0c\u867d\u7136 Cyber-Zero \u548c CTF-Dojo \u7684\u521d\u8877\u662f\u5e2e\u52a9\u5f00\u53d1\u8005\u548c\u7814\u7a76\u4eba\u5458\u5728\u8f6f\u4ef6\u90e8\u7f72\u524d\u53d1\u73b0\u5e76\u4fee\u590d\u6f0f\u6d1e\uff0c\u4f46\u540c\u6837\u7684\u80fd\u529b\u4e5f\u53ef\u80fd\u88ab\u6ee5\u7528\u4e8e\u8fdb\u653b\u76ee\u7684\uff0c\u6bd4\u5982\u81ea\u52a8\u5316\u53d1\u73b0\u5916\u90e8\u7cfb\u7edf\u7684\u6f0f\u6d1e\uff0c\u751a\u81f3\u5f00\u53d1\u6076\u610f\u5de5\u5177\u3002\u7279\u522b\u662f Cyber-Zero \u7684\u300c\u514d\u8fd0\u884c\u65f6\u300d\u65b9\u6cd5\uff0c\u964d\u4f4e\u4e86\u8bad\u7ec3\u9ad8\u6027\u80fd\u7f51\u7edc\u5b89\u5168\u667a\u80fd\u4f53\u7684\u95e8\u69db\uff0c\u4f7f\u5176\u66f4\u5bb9\u6613\u88ab\u66f4\u5e7f\u6cdb\u7684\u7fa4\u4f53\u83b7\u53d6\u548c\u4f7f\u7528\u3002\u8fd9\u79cd\u6c11\u4e3b\u5316\u7684\u8d8b\u52bf\u65e2\u610f\u5473\u7740\u5b89\u5168\u7814\u7a76\u7684\u666e\u53ca\uff0c\u4e5f\u610f\u5473\u7740\u98ce\u9669\u7684\u6269\u6563\u3002<\/p>\n\n\n\n

      \u5b9e\u9a8c\u7ed3\u679c\u5df2\u7ecf\u8bc1\u660e\uff0c\u57fa\u4e8e\u865a\u62df\u8f68\u8ff9\u6216\u6267\u884c\u9a8c\u8bc1\u6570\u636e\u8bad\u7ec3\u7684\u6a21\u578b\uff0c\u80fd\u591f\u5728\u591a\u4e2a\u57fa\u51c6\u4efb\u52a1\u4e0a\u8fbe\u5230\u63a5\u8fd1\u751a\u81f3\u5ab2\u7f8e\u95ed\u6e90\u524d\u6cbf\u6a21\u578b\u7684\u6027\u80fd\u3002\u8fd9\u8868\u660e\u5148\u8fdb\u7f51\u7edc\u5b89\u5168\u80fd\u529b\u7684\u6c11\u4e3b\u5316\u4e0d\u4ec5\u5728\u6280\u672f\u4e0a\u53ef\u884c\uff0c\u800c\u4e14\u6b63\u5728\u5feb\u901f\u5230\u6765\u3002\u5982\u4f55\u786e\u4fdd\u8fd9\u7c7b\u80fd\u529b\u66f4\u591a\u5730\u670d\u52a1\u4e8e\u9632\u5fa1\uff0c\u800c\u4e0d\u662f\u88ab\u6ee5\u7528\u4e8e\u653b\u51fb\uff0c\u5c06\u662f\u672a\u6765\u4e9f\u9700\u8ba8\u8bba\u7684\u8bae\u9898\u3002<\/p>\n\n\n\n

      \u5728\u672a\u6765\u7814\u7a76\u65b9\u5411\u4e0a\uff0c\u56e2\u961f\u63d0\u51fa\u4e86\u51e0\u4e2a\u503c\u5f97\u5173\u6ce8\u7684\u601d\u8def\u3002\u4e00\u4e2a\u662f\u6784\u5efa\u5b9e\u65f6\u66f4\u65b0\u7684 CTF \u57fa\u51c6<\/strong>\uff1a\u901a\u8fc7 CTF-Forge \u81ea\u52a8\u91cd\u5efa\u6bd4\u8d5b\u73af\u5883\uff0c\u628a\u6765\u81ea\u6d3b\u8dc3 CTF \u8d5b\u4e8b\u7684\u6311\u6218\u5bb9\u5668\u5316\uff0c\u7528\u4e8e\u52a8\u6001\u8bc4\u6d4b\u548c\u8f68\u8ff9\u91c7\u96c6\uff0c\u5b9e\u73b0\u53ef\u6269\u5c55\u3001\u5b9e\u65f6\u7684 benchmark\u3002\u53e6\u4e00\u4e2a\u65b9\u5411\u662f\u5f3a\u5316\u5b66\u4e60<\/strong>\uff0c\u5373\u8ba9\u7f51\u7edc\u5b89\u5168\u667a\u80fd\u4f53\u76f4\u63a5\u4e0e\u52a8\u6001\u73af\u5883\u4ea4\u4e92\uff0c\u5e76\u901a\u8fc7\u7ed3\u6784\u5316\u5956\u52b1\u83b7\u5f97\u53cd\u9988\u3002\u8fd9\u79cd\u8303\u5f0f\u6709\u671b\u7a81\u7834\u5355\u7eaf\u6a21\u4eff\u5b66\u4e60\u7684\u5c40\u9650\uff0c\u4f7f\u6a21\u578b\u80fd\u591f\u53d1\u5c55\u51fa\u66f4\u666e\u9002\u3001\u66f4\u5177\u9002\u5e94\u6027\u7684\u7b56\u7565\uff0c\u66f4\u597d\u5730\u5e94\u5bf9\u672a\u77e5\u7684\u5b89\u5168\u95ee\u9898\u3002<\/p>\n\n\n\n

      \u56e0\u6b64\uff0c\u672a\u6765\u7684\u5173\u952e\u5728\u4e8e\u5e73\u8861\u5f00\u653e\u4e0e\u5b89\u5168\u3002\u5728\u63a8\u52a8\u6280\u672f\u8fdb\u6b65\u4e0e\u666e\u53ca\u7684\u540c\u65f6\uff0c\u5efa\u7acb\u6709\u6548\u7684\u5b89\u5168\u62a4\u680f\uff0c\u9700\u7814\u7a76\u8005\u3001\u5f00\u53d1\u8005\u3001\u5b89\u5168\u673a\u6784\u4e0e\u653f\u7b56\u5236\u5b9a\u8005\u534f\u540c\u52aa\u529b\uff0c\u786e\u4fdd\u8fd9\u7c7b\u5f3a\u5927\u5de5\u5177\u4ee5\u8d1f\u8d23\u4efb\u7684\u65b9\u5f0f\u88ab\u5f00\u53d1\u548c\u4f7f\u7528\u3002\u552f\u6709\u5982\u6b64\uff0c\u624d\u80fd\u771f\u6b63\u589e\u5f3a\u6574\u4f53\u7f51\u7edc\u9632\u5fa1\u80fd\u529b\uff0c\u8fce\u63a5\u4e00\u4e2a\u66f4\u5b89\u5168\u7684\u667a\u80fd\u65f6\u4ee3\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

      \u5f53\u4eba\u4eec\u8fd8\u5728\u60ca\u53f9\u5927\u6a21\u578b\u80fd\u5199\u4ee3\u7801\u3001\u80fd\u81ea\u52a8\u5316\u529e\u516c\u65f6\uff0c\u5b83\u4eec\u6b63\u5728\u6084\u7136\u8e0f\u5165\u4e00\u4e2a\u66f4\u654f\u611f\u3001\u66f4\u5371\u9669\u7684\u9886\u57df \u2014\u2014\u7f51\u7edc\u5b89\u5168\u3002 \u60f3\u8c61\u4e00 […]<\/p>\n","protected":false},"author":1,"featured_media":1039,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[],"class_list":["post-1035","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-10"],"_links":{"self":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1035","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/comments?post=1035"}],"version-history":[{"count":1,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1035\/revisions"}],"predecessor-version":[{"id":1043,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1035\/revisions\/1043"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/media\/1039"}],"wp:attachment":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/media?parent=1035"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/categories?post=1035"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/tags?post=1035"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}