HaiAI123

Curated Global AI Tools Directory

SurgeIndustry update
↑ 85%

Cloudflare discloses cross-tenant KV cache leak on Workers AI

10/27/2026 — 10/27, 17:23·3 sources (all)·3 reports (all)

Key takeaways

  • Cloudflare disclosed a cross-tenant isolation incident on Workers AI: between August 27 and September 9, 2026, the GLM-5.3 model sent data f…

Latest turnCloudflare has disclosed a tenant-isolation incident on Workers AI: between August 27 and September 9, the GLM-5.3 model (@cf/zai-org/glm-5.3) mixed other customers' request data into responses 155 times, all on very long-context prompts, according to a report by OSChina. Security researcher Gadi Evron of Knostic reported the cross-request KV cache exposure through HackerOne on September 8.

Related AI tools

24-hour heatHeat index 190 · peak 216 · 4h ago
24 hours agonow

Reports on this story headlines open the original

Today
  1. Cloudflare has disclosed a tenant-isolation incident on Workers AI: between August 27 and September 9, the GLM-5.3 model (@cf/zai-org/glm-5.3) mixed other customers' request data into responses 155 times, all on very long-context prompts, according to a report by OSChina. Security researcher Gadi Evron of Knostic reported the cross-request KV cache exposure through HackerOne on September 8.

    开源中国AI score 66
  2. A Hacker News discussion flags a cross-request KV cache exposure issue in the GLM-5.3 Workers AI model, with the thread drawing 57 points and 155 responses. The reported flaw points to a potential cache-isolation risk in model serving, though the linked discussion provides limited technical detail.

    Hacker News · AIAI score 61
Yesterday
  1. SGLang released a security update to improve request isolation after Cloudflare notified it of cross-request KV cache exposure in its GLM-5.3 deployment on Workers AI on September 9, 2026. The two teams fixed the request-rejection path and strengthened cache-lifetime handling in disaggregated serving.

    LMSYS / SGLangFirst-partyAI score 50

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →