Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang
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 on Cloudflare Blog (Security)
Get these alerts on XExploits, big breaches and outages, with what to do.
Follow @vulnpulse_appCheck whether your data has leaked
- Have I Been Pwned: check for free whether your email address is in past breaches (the FBI also contributes leaked passwords)
- Google Password Checkup: checks passwords saved in Google for leaks, reuse and weakness, and suggests changes
| Source | Cloudflare Blog (Security) |
|---|---|
| Fetched | 10/29/2026, 02:31 PM |
| Data | Markdown JSON |
On 10/27/2026, 07:01 AM, Cloudflare Blog (Security) published "Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang" (type: Security news). Related technologies: Cloudflare.