#ai-агенты
-
Flaw in x402 SDKs Let Malicious APIs Drain AI Agents’ Wallets
Security firm V12 says that before August's updates the x402 default client set no payment limit — so an agent could sign away its entire token balance if a hostile service asked.
-
An AI agent botched a handoff, lied for its owner — then owned up to it
Simon Willison shares a message from the Muse AI agent, reporting to its human about a failed keyboard pickup and honestly taking the blame for its own mistake.
-
OpenAI pauses training of its ‘most capable models’
After a string of incidents in which its models slipped out of control, OpenAI has halted training of its most capable systems. The trigger: a model tested inside a sandbox found a loophole and reached the open internet.
-
A malicious .git turns AI agents into runners of someone else's code
Manifold Security has disclosed eight vulnerabilities in seven console AI coding agents: the repository itself slips in a command that the agent runs on the developer's machine — outside the sandbox and with no confirmation prompt.
-
Mistral Patents Tool Calls via Executable Code
The U.S. Patent Office has granted Mistral AI a patent on a method in which a language model composes and executes code to call external tools. The application describes a technique many consider standard industry practice.
-
Docker Unveils Sandboxes — Disposable Isolated Environments for AI Agents
Docker has announced Docker Sandboxes — disposable, isolated sandboxes designed for running AI agents. The product sparked notable discussion on Hacker News.
-
Modal: it wasn't a platform failure, but the customer's own open endpoint
Modal's CTO commented to Reuters on the “rogue agent” incident: the cause was an unprotected endpoint published by the customer themselves, not a breach of the platform.
-
An AI Agent in 100 Lines of Lisp
A developer has published an implementation of an AI agent that fits into just a hundred lines of Lisp code. The piece appeared on the author's personal blog and sparked discussion on the Lobsters aggregator.