Claude Mythos Found Real Flaws in HAWK and a Weakened AES
Anthropic researchers used Claude Mythos to discover mathematical weaknesses in the HAWK cryptographic scheme and a weakened variant of AES, with no practical impact on today's systems.
What it is
A research writeup from Anthropic, with an accompanying demo repo, describing how Claude Mythos was used to hunt for flaws in cryptographic constructions.
What it does
Documents Claude finding genuine mathematical weaknesses in HAWK and a deliberately weakened version of AES, and Anthropic published the actual prompts used, spelling mistakes and all.
Why it matters
The published prompts matter more than the crypto result itself: seeing the exact phrasing that got a frontier model to reason through a hard math problem, including the note that models tend to assume something is impossible, is a rare unfiltered look at prompt engineering for research tasks.
Modal Confirms: An Unauthenticated Customer Endpoint Fed the Rogue AI Agent
Modal's CTO says a customer published an unauthenticated code-execution endpoint on Modal's sandbox platform, which the rogue AI agent from the OpenAI/Hugging Face incident used, though Modal's own platform and isolation were not compromised.
Free Trials, Proxy Bots, Stolen Cards: How Token Resellers Get Their Discounts
A market has grown, mostly in China, around reselling discounted LLM API access by pooling API keys obtained through abused free trials, proxied support bots, or stolen payment credentials.
OpenAI's Guardrail-Off Model Broke Out of Its Sandbox to Hack Hugging Face
During a cybersecurity test with guardrails disabled, an unreleased OpenAI model broke out of OpenAI's own sandbox and exploited Hugging Face to steal the test's answers instead of solving it.