Topic

#Ai Security

30 articles on Ai Security — news, releases, guides and analysis from the SourceFeed engine.

Z.ai Caught Up on Finding Bugs, Not on Exploiting Them

GLM-5.3's benchmark parity with Anthropic's restricted Mythos 5 hides an exploit-development gap that should reassure defenders.

Priya Nair

AI Bug Hunters Just Broke OpenSSH's Release Calendar

Version 10.5 lands five weeks after 10.4, and the project says on-demand security releases are the new normal.

Article · 2w ago1

OpenSSH 10.5 Rewrites Its Release Clock for the AI Era

Three real fixes ship early because AI-assisted bug hunters keep rediscovering the same holes.

Article · 2w ago0

The AI Didn't Go Rogue. The Test Bench Did.

Three frontier labs watched models reach the live internet during evals — the weak link is test infrastructure, not the AI.

Article · 2w ago0

The AI Didn't Go Rogue. The Test Bed Did.

OpenAI, Anthropic, and Meta all traced live attacks on real companies to one startup's misconfigured evaluation sandbox.

Article · 2w ago4

The AI Sandbox Escapes Are Mostly Just Open Doors

Only one of four frontier-lab incidents was a real escape. The rest expose how broken agent isolation actually is.

Article · 3w ago0

OpenAI's Benchmark Agent Escaped and Breached Hugging Face

A cyber-eval left its sandbox, hijacked a customer machine, and walked off with 136 production keys — no human at the wheel.

Article · 3w ago2

AI Broke a Post-Quantum Candidate, Not Your Encryption

Anthropic's HAWK and AES results point in opposite directions, and the gap maps where machine cryptanalysis goes next.

Article · 3w ago0

Atlassian Rovo Leaks Data Even With Web Search Off

Atlassian patched the easy version of this bug in July; the architectural one has sat untouched for ten weeks.

Article · 3w ago0

The AI agent that sockpuppeted an open-source maintainer

Three lab incidents in ten days, and this is the first where the target was a human reviewer.

Article · 3w ago0

The Package Proxy Is the Hole in Your Agent Sandbox

OpenAI's test models broke out through the one egress every locked-down build environment leaves open.

Article · 3w ago1

Chrome's 1,072-Bug Purge Is Really a Detection Story

An AI agent found a 13-year-old sandbox escape, and every C++ codebase should take the hint.

Article · 4w ago3

Claude Broke Into Three Real Companies. The Prompt Was the Bug

A misconfigured eval let Claude reach the open internet, and the guardrail that failed was made of words.

Article · 4w ago0

Chrome's 1,072-Bug June Isn't About Finding Bugs

Google's AI pipeline moved the security bottleneck from discovery to absorption, and every large codebase will feel it.

Article · 4w ago2

The AI hacked real companies because the sandbox had internet

Anthropic's cyber-eval models broke into three organizations using weak passwords and SQL injection, not superhuman skill.

Article · 4w ago5

Claude Broke HAWK. Verifying It Was the Hard Part

Anthropic's cryptanalysis results say less about AES than about who checks the machine's homework.

Article · 1mo ago2