I Let Claude Opus Write a Chrome Exploit: The Next Model (Mythos?) Won't Need My Help?

Benchmarking GPT-5.6 Sol Medium, Sol Ultra, and Grok 4.5 on Chrome exploit development: after 2.096 billion tokens, only Sol Ultra produced a complete exploit chain.

The authors reverse-engineered Chrome's January 2026 fix for CVE-2026-0899, an out-of-bounds access in V8, built a trigger PoC, and defeated a stubborn CHECK to reach arbitrary read/write.

PortSwigger outlines its 2025 community plans: the growing PortSwigger Discord, and Burp on Tour stops including BSides Exeter, Hacking APIs Con NYC, SteelCon and RomHack.

Trail of Bits details Electron CVE-2025-55305: overwriting V8 heap snapshot files bypasses code integrity checks, enabling local backdoors in Signal, 1Password, Slack, and other Chromium-based apps.

XBOW's autonomous solve of a padding-oracle benchmark: the agent built an exploit to decrypt a server cookie and bypass authentication — an attack the author calls the hardest he taught at NYU.

ExploitBench measures how far AI climbs a five-tier exploitation ladder from crash to full code execution; private model Mythos matched trained specialists on V8 targets.