[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$flp7Wt1eTbhbr4geCK2bAT3314pf0ZznlwdTvciVpNpc":3},{"lesson":4},{"id":5,"slug":6,"article_id":7,"title":8,"body":9,"prevention":10,"framework_refs":11,"framework_notes":17,"status":23,"created_at":24,"published_at":25,"article":26,"tags":30,"podcasts":43,"controls":44},"8f9ab96f-2918-4251-a059-9a425499bbe1","unpatched-rce-in-open-source-ai-framework-leads-to-data-breach-across-users-and-ai-agents","2bdd800a-3268-4357-8ff9-61b702e41e33","Unpatched RCE in Open-Source AI Framework Leads to Data Breach Across Users and AI Agents","NEAR AI's inference server was compromised because it ran a vulnerable version of SGLang, an open-source AI inference engine with a known remote code execution flaw. The attacker leveraged this flaw to execute arbitrary code on the server, exposing personal data for roughly 100 users and sensitive logs—including API keys—from 18 AI agents. This incident illustrates the risk of deploying open-source dependencies without a disciplined process for tracking upstream vulnerability disclosures and applying patches promptly. The exposure of API keys in logs further compounded the breach, turning a single point of failure into a credential compromise event. Organizations building on open-source AI tooling must treat third-party inference engines with the same rigor as any other production dependency.","**Immediate actions:**\n- Identify and upgrade all instances of SGLang (and similar open-source inference engines) to the patched version immediately.\n- Rotate all API keys and credentials that were accessible on affected servers, even if exposure is unconfirmed.\n- Audit application and server logs to ensure secrets and credentials are never written in plaintext.\n\n**Long-term improvements:**\n- Establish a formal software inventory (SBOM) for all open-source components used in AI\u002FML pipelines so vulnerabilities can be rapidly identified and scoped.\n- Subscribe to upstream security advisories for every open-source dependency and define SLA-based patching windows for critical CVEs.\n- Implement network segmentation to isolate inference servers from each other and from internal systems, limiting blast radius of any single server compromise.\n\n**Detection measures:**\n- Deploy runtime monitoring and anomaly detection on inference servers to flag unexpected code execution or unusual process spawning.\n- Implement secrets-scanning in CI\u002FCD pipelines and log aggregation systems to detect and alert on credential exposure before an attacker finds them.\n- Conduct regular vulnerability scans of internet-facing AI infrastructure, including third-party inference engines.",[12,13,14,15,16],"CIS Control 2: Inventory and Control of Software Assets (v8)","CIS Control 7: Continuous Vulnerability Management (v8)","NIST SP 800-53 SI-2: Flaw Remediation","NIS2 Article 21(2)(e)","OWASP A06:2021 Vulnerable and Outdated Components",{"NIS2 Article 21(2)(e)":18,"NIST SP 800-53 SI-2: Flaw Remediation":19,"OWASP A06:2021 Vulnerable and Outdated Components":20,"CIS Control 7: Continuous Vulnerability Management (v8)":21,"CIS Control 2: Inventory and Control of Software Assets (v8)":22},"This point requires security in acquisition, development, and maintenance including vulnerability handling, which directly covers the obligation to track and patch known vulnerabilities in dependencies like SGLang.","SI-2 requires organizations to identify, report, and correct information system flaws; applying this control would have mandated patching the SGLang vulnerability promptly after disclosure.","The breach occurred precisely because NEAR AI ran an outdated, vulnerable version of an open-source component, which is the exact risk scenario OWASP A06 addresses.","A continuous vulnerability management process—including subscription to upstream advisories and timely patching—would have ensured the SGLang RCE flaw was remediated before it could be exploited.","Maintaining an accurate inventory of open-source software components, including SGLang, would have enabled NEAR AI to quickly identify and scope the vulnerable deployment when the RCE flaw was disclosed.","published","2026-11-10T20:21:39.656998+00:00","2026-11-10T20:21:39.641+00:00",{"id":7,"url":27,"slug":28,"title":29},"https:\u002F\u002Fthehackernews.com\u002F2026\u002F11\u002Fsglang-rce-exploit-breaches-near-ai.html","sglang-rce-exploit-breaches-near-ai-steals-data-from-18-agents-and-about-100-use-12e105","SGLang RCE Exploit Breaches NEAR AI, Steals Data From 18 Agents and About 100 Users",[31,37],{"id":32,"name":33,"slug":34,"description":35,"color":36},"05757c8d-6b93-4194-b35d-7359e7d33b0e","Vulnerability Management","vulnerability-management","Missing scans, no risk prioritization","#fb923c",{"id":38,"name":39,"slug":40,"description":41,"color":42},"af7fce9e-1ce8-4156-93bc-09dcfbfdf29d","Patch Management","patch-management","Unpatched vulnerabilities, delayed updates","#ef4444",[],[45,48,52],{"slug":34,"name":46,"problem":47},"Vulnerability scanning and prioritisation","You know which systems are vulnerable, and which ones to fix first.",{"slug":49,"name":50,"problem":51},"immutable-backup","Immutable backup and recovery","Ransomware cannot delete or encrypt your way back.",{"slug":53,"name":54,"problem":55},"incident-response-retainer","Incident response retainer","Experts are on call before the worst day, not searched for during it."]