Memory corruption

Security briefings tagged with Memory corruption

Needs attention

Active incidents, new guidance or unresolved scope that security teams should review now.

Last checked

Patch

Linux IPv4 TCP use-after-free

CVE-2026-98432

A race condition in Linux's IPv4 TCP connection handling can expose kernel memory, crash a host, or allow remote code execution before application authentication.

Why this needs attention

An unauthenticated remote attacker can reach the vulnerable Linux kernel code before application authentication. Researchers demonstrated root command execution against MPTCP listeners, and distribution-specific patch availability requires active review. Exploitation in the wild is not established in the reviewed sources.

What to do now

Check the running host kernel against your distribution's CVE advisory, apply its fixed kernel, and complete the required reboot or vendor-supported activation procedure. Verify that the fix is active, including on container hosts and Kubernetes nodes.

Last checked

Patch

php2shell

CVE-2026-109618 · CVE-2026-109619 · CVE-2026-115560

Two PHP runtime flaws can allow unauthenticated code execution, data exposure, or crashes. A related follow-up, CVE-2026-115560, adds a configuration-dependent session upload progress RCE vector.

Why this needs attention

CISA has listed CVE-2026-109618 and CVE-2026-109619 as known exploited. The related CVE-2026-115560 requires a separate session-configuration assessment; exploitation of that follow-up is not established in the reviewed sources. These PHP runtime issues require attention beyond application-code or plugin updates.

What to do now

Patch the PHP runtime using the original security advisory and separately review PHP's session upload progress advisory for CVE-2026-115560. Keep compensating controls in place while patching, and ask managed hosting providers to confirm each issue's status.

Last checked

Assess exposure

node2shell

CVE-2026-111497

CVE-2026-111497 is a memory-corruption flaw in the shared Brotli decoder. Any service that decompresses attacker-controlled Brotli, in a request or an HTTP response, can be exposed.

Why this needs attention

The risk extends beyond Node.js deployments, and affected runtime and server-side fetch paths require active assessment. Active exploitation is not established in the reviewed sources, but scope and remediation guidance are still developing.

What to do now

Identify every Brotli decoder in running runtimes, dependencies, and container images, then use each vendor's security guidance to update or otherwise remove the vulnerable decoder.