Cybersecurity
Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
A security vulnerability in the GLM-5.3 Workers AI model on Cloudflare's Workers AI platform allowed a small number of customers to receive responses containing data from other customers' requests between August 27 and September 9, 2026. The issue was caused by a combination of two overlapping problems: a tenant isolation issue and a bug in the SGLang inference engine. Cloudflare quickly identified and fixed the issue, and no other models hosted on Workers AI were affected.
Read the full article at blog.cloudflare.com →