The company says the attacker could have seen requests on one inference server, including ones running in confidential mode.
By [Emerson Redding](https://runtimewire.com/author/emerson-redding)
· Published
Primary source: [near.ai](https://near.ai/blog/security-incident-report-sglang-vulnerability)
Why it matters #
NEAR markets confidential inference as protection for user data, and its report says a flaw in the open-source inference engine reached requests inside that environment.
NEAR AI said an attacker exploited a vulnerability in the open-source SGLang inference engine to run code on one of its servers and gain access to its infrastructure.
In a Nov. 9th incident report, the company said it had fixed the vulnerability on its servers and reported it to the SGLang project. The report did not identify the flaw, confirm an upstream fix or establish whether a fix was available to other SGLang operators.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.