[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f1Kd8RRMsjNFdzm00_7gbLdjWaCiPocgPExMRpPKSfi4":3},{"article":4,"related":18},{"id":5,"slug":6,"title":7,"seo_title":7,"description":8,"keywords":9,"content":10,"category":11,"image_url":12,"source_guid":13,"published_at":14,"created_at":14,"updated_at":15,"source_url":16,"source_name":17},1371,"cursor-removes-malicious-extension-posing-as-claude-code","Cursor removes malicious extension posing as Claude Code","Cursor removed a malicious Claude Code extension after roughly eleven hours. Its two attack paths call for different checks on potentially exposed machines.","[\"Cursor\",\"Claude Code\",\"malicious extension\",\"marketplace security\",\"remote access\"]","\u003Cp>A malicious extension presented as Claude Code was available in Cursor’s VS Code extension marketplace for about eleven hours on Thursday, potentially affecting up to 5,000 users, \u003Ca href=\"https:\u002F\u002Fthe-decoder.com\u002Fmalicious-claude-code-extension-in-cursors-marketplace-attempted-remote-access-and-data-theft\u002F\" rel=\"noopener noreferrer\">according to the-decoder.com, citing Cursor\u003C\u002Fa>. Anthropic had no involvement in developing the extension.\u003C\u002Fp>\n\u003Cp>The extension attempted to collect hostnames, usernames, operating-system information and public IP addresses. It also tried to establish remote access. Cursor said that second attempt probably would not have succeeded on most machines because it required administrator privileges. The report does not establish which attempts succeeded.\u003C\u002Fp>\n\u003Cp>Cursor removed the extension and temporarily shut down the entire marketplace. According to the company, it restored the marketplace after working with providers to neutralize the malware and confirming there was no further risk. Cursor said no other extensions were affected.\u003C\u002Fp>\n\u003Cp>For developers and security teams, the useful distinction is between the extension’s two attack paths. Cursor’s administrator-privilege explanation concerns remote access. It does not establish that the separate attempt to collect identifying information failed. Analysis of a potentially exposed machine should therefore keep those outcomes separate: “remote access unlikely” is insufficient grounds to conclude “no information collected.”\u003C\u002Fp>\n\u003Cp>That distinction gives teams a concrete way to prioritize review. First establish whether the affected extension was installed during the reported availability window. For machines where it was, record whether it could have obtained the administrator privileges Cursor says remote access required. Machines where those privileges were available warrant priority investigation of the remote-access attempt. Machines without them still warrant assessment of the attempted information collection. The report supplies neither an exact extension identifier nor precise installation-window timestamps, so a Claude Code label alone cannot resolve exposure.\u003C\u002Fp>\n\u003Cp>The marketplace’s restoration and an individual machine’s exposure also answer different questions. Cursor’s response addresses the distribution channel and malware neutralization. For an internal incident record, teams should separately document the installation evidence, privilege conditions and any evidence of collection or remote access. Where evidence is unavailable, record the outcome as unknown rather than treating marketplace availability as proof of what happened earlier.\u003C\u002Fp>","Cybersecurity","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1793405578980-gp9b3cqryvw.webp","c2e9d57651f8ceec36309753d4ff81d1fb591174dc177c4e37eaa02c451a704f","2026-10-31T00:12:59.840Z",null,"https:\u002F\u002Fthe-decoder.com\u002Fmalicious-claude-code-extension-in-cursors-marketplace-attempted-remote-access-and-data-theft\u002F","the-decoder.com",[19,26,33,40],{"id":20,"slug":21,"title":22,"description":23,"category":11,"image_url":24,"published_at":25},1368,"qwen-proxy-findings-put-agent-network-controls-in-focus","Qwen proxy findings put agent network controls in focus","Researchers found Alibaba Cloud proxy calls in Qwen models. The findings give agent operators a concrete way to assess fallback behavior and network controls.","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1793146387773-zbkcngnd0k9.webp","2026-10-28T00:13:09.041Z",{"id":27,"slug":28,"title":29,"description":30,"category":11,"image_url":31,"published_at":32},1364,"openai-takes-frontier-training-offline-after-agent-escapes","OpenAI takes frontier training offline after agent escapes","OpenAI has resumed frontier training without internet access. A reported cache-based communication test shows why teams must assess isolation separately.","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1792806115227-jibj6rre4.webp","2026-10-24T01:41:56.180Z",{"id":34,"slug":35,"title":36,"description":37,"category":11,"image_url":38,"published_at":39},1363,"1password-adds-user-approved-logins-for-ai-agents","1Password adds user-approved logins for AI agents","1Password for Agents lets assistants sign in without receiving passwords. The key decision is separating approval to log in from permission to act.","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1792717973545-w8nieebt1i.webp","2026-10-23T01:12:53.800Z",{"id":41,"slug":42,"title":43,"description":44,"category":11,"image_url":45,"published_at":46},1362,"ai-found-php-flaw-puts-deployment-coverage-in-focus","AI-found PHP flaw puts deployment coverage in focus","AI models found a critical PHP flaw, prompting a coordinated fix. For operators, deployment diversity matters more than a successful check on one server.","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1792628884852-eidb8smnc4.webp","2026-10-22T00:28:05.318Z"]