News · DiggingBeagle record

A Node2Shell attempt reached a sensor; laboratory RCE is a different observation

ELLIO reported a crafted Brotli request targeting Node.js 24.22.0 on October 26 and separately showed exploit effects in a lab. The sensor attempt is not proof of a successful production intrusion.

Published
Oct 27, 2026

The report

On October 26, ELLIO reported detecting a specially constructed Brotli payload aimed at Node.js 24.22.0. The captured attempt was approximately 23.07 MB. The researchers then described a separate laboratory reproduction involving a stock Node executable and an x86-64 return-oriented chain. The two results have different evidentiary meanings: one establishes an observed attempt at a sensor; the other supports exploitability under tested conditions.

From a decoder defect to an execution path

The underlying issue, CVE-2026-111497, involves unsafe behavior in Brotli decoding. Node.js published corrected releases on October 21, including 22.23.4, 24.22.1 and 26.12.1. ELLIO’s sample description identifies the targeted pre-patch build and reports a sequence of exploit stages. It does not provide a verified production-victim trace showing that the payload completed code execution.

The bug is notable in the Digging Beagle corpus because AI-assisted security research was credited in its discovery history. That attribution concerns how the defect was found, not an assertion that the later attacker used AI. The canonical Brotli case keeps discovery, observed exploit attempts and application-specific effects separate.

One library, several outcomes

JFrog’s follow-up describes different consequences where Brotli appears inside ClickHouse, FFmpeg/JPEG XL, .NET and Arrow. Some reproduced outcomes are crashes, memory corruption indicators or sustained CPU consumption. Those outcomes do not automatically become remote code execution just because a different Node.js harness demonstrated a control-flow exploit. Attack surface and embedded library versions matter as much as the CVE label.

Organizations using Node should check the fixed runtime builds and any statically bundled copies of vulnerable decoder code. A package manager reporting a patched high-level dependency is not, by itself, proof that all embedded native copies were replaced.

Evidence boundary: an incoming attack sample plus a controlled exploit reproduction supports serious urgency. It does not establish a number of compromised hosts or successful exploitation of every application that embeds Brotli.

Research behind this