TechCloudflare Blog1h ago
Cross-request KV cache exposure in the GLM-5.3 Workers AI model
Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang
TL;DRBug allowed AI model requests to leak cached data between different users.
Why it matters: Cache leaks in shared AI services can expose sensitive prompts and responses across customers.
A configuration mismatch in GLM 5.3 on Workers AI triggered an SGLang bug that exposed KV cache pages across requests. Here is how we addressed the issue.
Read full articleSource: Cloudflare Blog · Opens in new tab