Tag: trusted execution environment
2 articles

SGLang Flaw Exploited, Breaches NEAR AI Systems, Data Stolen
A security breach hit NEAR AI when an attacker exploited a flaw in the SGLang inference engine, allowing them to run code on an inference server and steal data from 18 AI agents and sensitive account details of around 100 users. The hack occurred despite the server running in a secure, confidential mode within a trusted execution environment.

Confidential Computing Flaws Expose Trust Risks
Researchers have uncovered alarming flaws in confidential computing, including a high-severity vulnerability rated 7.5, which can trick cryptographic systems into verifying the wrong machine, putting trust at risk. This weakness was found in protocols like attested TLS, which failed to ensure sensitive data reaches its intended destination.