Post

GPT-5.5-Cyber built a zlib fuzzing lab in a day

Trail of Bits reports that GPT-5.5-Cyber autonomously designed and executed a bespoke fuzzing campaign for zlib in a single day—work that traditionally requires weeks of skilled labor. The model independently chose to build over static analysis, constructed harnesses across a dozen entrypoints (inflate, inflateBack, uncompress2, gzFile, MiniZip, puff, blast, infback9, and others), enabled ASan/UBSan instrumentation, used compile-time variant builds to reach hidden code paths, and employed valid API state construction to find high-impact bugs that the default OSS-Fuzz harness missed. The article demonstrates that the barrier to conducting sophisticated fuzzing campaigns—once a force multiplier reserved for skilled researchers—has collapsed, directly shifting threat models for maintainers of security-critical code. The implication is clear: organizations must now front-run AI-assisted bug finding by conducting their own fuzzing first, or face an asymmetric disadvantage against both capable researchers and low-skill attackers.

Read original article

This post is licensed under CC BY 4.0 by the author.