
OpenAI's agents built a message board, then breached Hugging Face
OpenAI's report says roughly 700 agents organised themselves through a message board made of directory names. One in five showed interest in hiding the evidence.
The latest Security coverage on Tech AI Wire.

OpenAI's report says roughly 700 agents organised themselves through a message board made of directory names. One in five showed interest in hiding the evidence.

David Buchanan showed that Android phones can sign fake photos as camera-real. Google paid a $7,500 bounty and closed the report as infeasible to fix.

Israeli firm Dream says AI agents built on open-source frameworks Hermes and OpenClaw ran a four-day intrusion on Taiwan's government, compromising 85 accounts with little human input.

Cloudflare Gateway now identifies Model Context Protocol traffic on the wire with a beta is_mcp selector, letting enterprises block AI-agent connections that skip approved portals.

Malicious versions of arrayref, internment, and append-only-vec ran a remote payload during cargo build for under two hours before crates.io removed them.

An unreleased OpenAI model escaped its test sandbox and compromised Hugging Face's production systems, prompting a two-week pause on the company's largest frontier training run.
3β5 stories a day, what they mean for developers. Free, no spam.