NEAR AI Breach Exposes Data of 100 Users and 18 Agents
An attacker exploited a SGLang flaw to access NEAR AI servers, stealing user details and agent data.
An attacker exploited a remote code execution flaw in SGLang to run code on one of NEAR AI's inference servers last Sunday. The breach allowed the attacker to access an internal system behind Private Chat and download account details for about 100 users. The stolen data included email addresses, names, and for some accounts, plan and balance information. NEAR AI stated that it found no evidence that stored Private Chat conversations or uploaded files were accessed.
The attacker also accessed 18 AI agents on an older deployment that did not run inside a trusted execution environment. Data taken from these agents included chat history, settings, and logs. Additionally, two customers' NEAR AI Cloud API keys were exposed in logs during the breach. NEAR AI noted that its TEE does not protect against vulnerabilities in the software running inside it.
Following detection, NEAR AI revoked all active sessions, administrator credentials, and the exposed API keys. The company replaced all agent tokens, took Private Chat offline, and deployed a fix for the SGLang flaw on all inference servers. NEAR AI completed a full forensic review and contacted affected users via [email protected]. The company warned users to be careful with unsolicited messages about their account or balance.
Sources
- The Hacker NewsSGLang RCE Exploit Breaches NEAR AI, Steals Data From 18 Agents and About 100 Users ↗
- DecryptNear AI Breach Exposes User Details and Data From 18 AI Agents ↗
Every sentence of this story was checked against facts quoted from these articles. How we make the news
Spotted an error? Tell us
