discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.

Cloudflare Workers Spectre Attack Leaks JWT at Up to 12 Bits/Second

Researchers disclose a Spectre attack against Cloudflare Workers that leaked JSON Web Token (JWT) from a co-located Worker process, with mitigation measures in place.

By Swati Khandelwal·Aug 19·thehackernews.com·1 min read

Intelligence analysis by Qwen 2.5 (3B)

Cloudflare Workers Spectre Attack Leaks JWT at Up to 12 Bits/Second
Image: thehackernews.com

Cybersecurity researchers have disclosed details of a remote Spectre attack against Cloudflare Workers that leaked a JWT from a co-located Worker at up to 12 bits per second. Mitigation measures are already deployed by Cloudflare.

Why it matters

This attack highlights the importance of robust security protocols in cloud environments, especially for services like Cloudflare Workers which run code from multiple tenants within the same process.

A Spectre attack let someone peek at a secret code snippet from another computer program running on the same server. Cloudflare fixed this by making sure different programs can't see each other's secrets.

Analysis

{"#shared-worker-leakage":"The Spectre attack against Cloudflare Workers involved a JWT leak at up to 12 bits per second. This was achieved by co-locating an attacker Worker and a victim Worker, with the JWT intentionally placed in the victim's memory.","#detection-mechanisms":"Cloudflare implemented DyPrIs (Dynamic Process Isolation) as a defense mechanism against Spectre attacks. However, researchers found that WebSocket communications could provide a remote timing source, leading to reduced detection signals and increased leakage rates.","#mitigation-strategies":"To mitigate the attack, Cloudflare improved DyPrIs, integrated V8 Sandbox, and deployed Memory Protection Keys (MPK)-based in-process isolation. These measures include restricting local timers, isolating suspicious scripts into a separate process, and using hardware-enforced protection keys."}

Key points

  • Cloudflare Workers experienced a Spectre attack leaking JWT at up to 12 bits per second
  • Mitigation measures were already deployed by Cloudflare including DyPrIs, V8 Sandbox, and MPK-based in-process isolation
  • The attack required co-located Worker processes and intentional placement of the JWT
The Upside

The mitigation strategies deployed by Cloudflare have likely prevented any real damage from occurring, and they continue to improve their security measures based on research findings.

The Downside

While the attack was mitigated, it shows that even with advanced security protocols in place, there is still room for improvement. The Spectre family of attacks remains a threat.

Originally reported at

thehackernews.com

Discernion covers the story. Read the full piece at the source.

Tagscloudflare-workersspectre-attackjwt-leakagedynamic-isolationmemory-protection-keys

Author

Swati Khandelwal

Intelligence analysis by

Qwen 2.5 (3B)

Published

Aug 19, 2026

Source

thehackernews.com

Share

Topics

cloudflare-workersspectre-attackjwt-leakagedynamic-isolationmemory-protection-keys

Related

More from this desk

Aug 20·bleepingcomputer.com

OpenAI confirms ChatGPT is down as logins and signups fail

OpenAI's ChatGPT is experiencing a major outage, affecting users worldwide. Users are unable to sign in, create accounts, or load chats, including previous conversations.

Aug 19·bleepingcomputer.com

Sakura Internet hack exposes data of up to 1.36 million accounts

Japanese cloud provider Sakura Internet disclosed a breach that may have impacted up to 1.36 million member accounts.

Aug 19·thehackernews.com

OpenAI Pauses Frontier RL Training as It Tightens Defenses Against Unsafe AI Behavior

OpenAI pauses reinforcement learning training for its latest AI models to strengthen defenses and monitor behavior.

Aug 19·bleepingcomputer.com

US Warns of AI-Powered Attacks on Siemens PLCs in Critical Infrastructure

U.S. agencies warn of AI-driven attacks targeting Siemens PLCs in critical infrastructure, including manufacturing, energy, and water systems.