Calif Reports OpenJDK Base64 Flaw Found With AI Agent; Oracle Releases Patch

Summary
Calif says its researcher used OpenAI's Codex to find CVE-2026-114938, a memory-corruption flaw in OpenJDK's Base64 decoder that may expose memory or enable remote code execution. Oracle has released a patch.
Key points
- Calif identified the flaw in OpenJDK's Base64 decoder and calls it Java4Shell.
- The company says malformed input—64 “A” characters followed by “=”—can trigger memory corruption.
- Calif says the flaw may expose sensitive memory or allow remote code execution; exploitability depends on how an organization uses Base64, and the company says there is no universal exploit.
- Calif researcher Quang Luoung directed OpenAI's Codex coding agent during the audit; the company says discovery and exploit development took three hours and cost $200 in tokens.
- Oracle released a patch after Calif's disclosure.
- Security expert Kevin Beaumont advised organizations to patch routinely and disputed descriptions of the flaw as world-ending.
Article Details
- Vulnerability Types
- Memory corruption
- Severity
- Critical
- Exploitation Status
- unknown
- Exploit Availability
- claimed_or_private
- Patch Status
- available
CVE
People
Vendors
OpenAIOpenJDK and chose to look at the platform's compiler, cryptography and APIs. Luoung directed Codex, OpenAI's coding agent, to look at each target "like any serious security researcher in 2026." The company didOracleSince Calif's disclosure, Oracle said it released a patch for the vulnerability.
Products
JavaJust weeks after the discovery of PHP2Shell, artificial intelligence models found another critical vulnerability, this time in Java.OpenJDKCybersecurity startup Calif said in a blog post Tuesday that it discovered CVE-2026-114938, a memory-corruption vulnerability in OpenJDK's Base64 decoder, a popular Java runtime platform. It calls the flaw Java4Shell.
