# Anthropic's unreleased Claude AI finds critical Linux kernel flaw CVE-2026-98432

Digest AI · Research · published 2026-10-30T11:55:00Z

Canonical: https://digestai.news/story/anthropic-s-unreleased-claude-ai-finds-critical-linux-kernel-flaw-cve

## Summary

Anthropic says an unreleased version of its Claude AI model discovered a critical Linux kernel vulnerability, tracked as CVE-2026-98432, with a severity score of 9.8 out of 10. The flaw is a use-after-free memory error in the networking code that handles Multipath TCP (MPTCP) connections. Anthropic demonstrated a proof-of-concept exploit that achieves root access in about 53 minutes by leaking kernel memory, bypassing KASLR, hijacking control flow, and executing arbitrary shell commands. No user interaction is required.

The vulnerability affects eight Linux kernel branches dating back to version 4.4. Fixed versions include 5.10.272, 5.15.223, 6.6.159, 6.1.190, 6.12.113, 6.18.56, 7.2.10, and 7.3-rc7. Debian, Oracle Linux, SUSE, and Amazon Linux have released patched packages for some versions. Ubuntu and Red Hat Enterprise Linux had not shipped fixes at the time of writing. Administrators are advised to check kernel versions, apply updates when available, and verify whether MPTCP listeners are enabled on their servers.

## Key points

- Unreleased Claude AI found CVE-2026-98432, a 9.8 severity Linux kernel use-after-free flaw
- Proof-of-concept exploit gains root access in ~53 minutes without user interaction
- Eight kernel branches back to v4.4 affected; patches out for some distros, not Ubuntu or RHEL yet

## Why it matters

An AI system autonomously discovered and exploited a critical zero-day vulnerability, demonstrating both defensive potential and offensive risk for infrastructure security.

## Sources

1. [Anthropic's Claude AI found a critical Linux flaw that could give hackers root access](https://tech.yahoo.com/ai/claude/articles/anthropics-claude-ai-found-critical-185522162.html) (tech.yahoo.com, 2026-10-30)

## Cite

Digest AI, "Anthropic's unreleased Claude AI finds critical Linux kernel flaw CVE-2026-98432", 30 October 2026, https://digestai.news/story/anthropic-s-unreleased-claude-ai-finds-critical-linux-kernel-flaw-cve

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/anthropic-s-unreleased-claude-ai-finds-critical-linux-kernel-flaw-cve.json