Google Brotli RCE hits Node.js, Python, Java, .NET
This isn't a single-runtime patch drill; Brotli sits underneath enough HTTP handling to turn inventory into incident response.
TL;DR
CIS says Google Brotli 1.0.3 through 1.2.0 carries an out-of-bounds write flaw, CVE-2026-111497, with public proof-of-concept code and third-party reports of exploitation attempts as of Oct. 27. Contractors, primes, subs, and ISPs should check Node.js 22.x, 24.x, and 26.x below fixed releases, plus Python Brotli, Java brotli4j, .NET Brotli decompression, Bun, Envoy, Istio, and bundled Linux packages. End-of-life Node.js 20.x, 23.x, and 25.x won't get fixes.
CIS's advisory turns Brotli from a quiet compression dependency into a Monday-morning inventory problem. The vulnerable range is Google Brotli 1.0.3 through 1.2.0, and the reported impact is not a polite denial-of-service bug: an unauthenticated remote attacker can feed crafted Brotli-compressed data to a vulnerable decoder, corrupt memory, disclose sensitive data, crash the process, or potentially execute code as that process. CIS initially reported no wild exploitation, then updated the advisory on Oct. 27 to say public proof-of-concept code is available and third parties have reported exploitation attempts.
The uncomfortable part is the dependency map. This is not just "update Node" and move on. CIS lists Node.js before 22.23.4, 24.22.1, and 26.12.1, with no fix coming for end-of-life 20.x, 23.x, and 25.x. It also lists Bun 1.4.3 and earlier, Python Brotli 1.2.0 and earlier, brotlicffi 1.2.0.x and earlier, Java brotli4j before 1.24.0, Microsoft .NET applications that enable or explicitly use Brotli decompression, Envoy Proxy and Istio when the Brotli decompressor is configured, and other packages that bundle or link the vulnerable library.
For federal contractors and service providers, the practical question is where attacker-controlled compressed data reaches decompression code. CIS names compressed HTTP request bodies, Node.js fetch responses with Content-Encoding: br, and compressed files as exposure paths. That means the owner may be the web platform team, the API gateway team, or the application group that pulled in a language binding three years ago and forgot it existed. The SBOM people get to be useful here, but only if the runtime owners can tie component names to deployed services.
There are still gaps. CIS says .NET Framework is not affected, while .NET applications using Brotli decompression are in scope. It gives fixed releases for Envoy and Istio, but the advisory's broader category of "other software or Linux distribution packages" remains deliberately open-ended. That's probably correct, because compression libraries get bundled in places asset inventories don't model well. It is also the reason this one deserves a faster queue than the usual vendor-advisory rhythm: proof of concept plus reported exploitation attempts changes the job from tracking a CVE to finding every decoder that will accept untrusted Brotli input.
Published ·Deep Fathom