Vulnerability / research
AI-assisted Java4Shell research exposes an OpenJDK decoder weakness
Reported
Published here
What happened
Calif reports using Codex to find a memory-corruption flaw in OpenJDK’s accelerated Base64 decoder and directing AI agents to develop application-specific exploit chains. The researchers supplied heap-analysis tooling and exploitation strategies.
Parties and reporting
Where this happened
Victim location not disclosed. Real organizations or systems were affected or targeted, but the sources do not say where. Explore the world map.
AI tools and systems: OpenAI Codex, OpenJDK / CVE-2026-114938
Outcome
Calif demonstrated memory disclosure and code execution in controlled applications. Oracle and OpenJDK issued fixes on November 17. Google documented interim mitigations and a continuing managed-service update rollout.
Operating lesson
Check runtime and CPU prerequisites, apply the runtime vendor’s mitigation, and verify rebuilt images and restarted workloads use corrected configurations.
Evidence and limits
The affected path requires suitable x86-64 AVX-512 VBMI hardware; ARM64 is excluded. Exploits require application-specific conditions. AI assistance is the researcher’s account; no malicious AI use or customer compromise is established here. Included as research, not an AI-product CVE.