October 1, 2026
- Anthropic's Frontier Red Team says open-weight GLM-5.3 (Zhipu/Z.ai) is close to frontier at exploit development. On ExploitBench, which targets known Chrome V8 bugs, GLM-5.3 produced a working exploit in 50 of 410 attempts versus 56 for Claude Mythos Preview; on Anthropic's internal OSS-Fuzz-derived binary exploitation benchmark it took full control of the target in 4% of tasks versus 6%. Paired with one human expert for a day, it found previously unknown bugs in a widely used browser's JS engine and chained them into a page that read a private SSH key; the smaller Flash variant weaponized a freshly disclosed Chrome bug for $20.40 in API costs. Safeguards strip out easily and unlocked copies circulate; US CAISI corroborates the findings (The Decoder). @emollick notes Anthropic's incentives but calls the trajectory — same capabilities, no guardrails — hard to dispute.
· AI & Model Security
in An Encoded Character Away From Admin: Cisco SD-WAN Manager Exploited in the Wild
July 29, 2026
- Anthropic's Claude Mythos Preview found weaknesses in real cryptographic algorithms, including an improved attack on HAWK — a post-quantum signature scheme humans had scrutinized for two-plus years — cracked in ~60 hours at about $100K of API cost, plus a faster attack on round-reduced AES. Anthropic stresses nothing in production is affected today, but it's a signal about AI cryptanalysis (Anthropic, The Hacker News). Bruce Schneier's write-up frames the broader benchmark question of how well LLMs can actually do cryptanalysis (Schneier).
· AI & Model Security
in Artifactory Zero-Days Confirmed as the Hugging Face AI Agent's Escape Route
June 19, 2026
- Squidbleed (CVE-2026-47729) is a 29-year-old Heartbleed-style memory disclosure affecting every version of Squid Proxy in its default config — notable also because it was surfaced by Anthropic's Claude Mythos Preview code-auditing (calif.io).
· Vulnerabilities & Exploits
in FortiBleed Burns 70,000+ Fortinet Edge Devices While a Leaked GitHub Token Guts Novo Nordisk