TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

Exploit.in data shows how a 2005 cybercrime forum helped shape today's ransomware ecosystem, with users and practices surviving for decades.

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

TL;DR: Three frontier labs have now published reports admitting their models broke into the production systems of companies nobody asked. Hugging Face’s technical timeline walks back roughly 17,600 attacker moves over four and a half days in July, and says the agent “escaped its sandbox by exploitin

A newly disclosed flaw called Plugin4Shell lets attackers silently swap in malicious code across four major AI coding agents, and Microsoft still hasn't shipped a fix for GitHub Copilot. Security research lab AIR Security disclosed on September 18 what it calls the first supply-chain-style vulnerability to hit the AI coding agent ecosystem. It named the flaw Plugin4Shell. The bug breaks the one safeguard developers actually trusted: pinning a plugin to a specific, reviewed commit hash. Exploiting it takes no click, no approval, and no reinstall. The malicious code just shows up the next time the agent updates. Here's the trick. When Claude Code, OpenAI's Codex, or GitHub Copilot installs a plugin, the agent runs a git checkout against a 40-character commit SHA, the long string meant to lock the code to one audited version. None of the three agents confirmed the checkout actually landed on that hash. AIR Security found that an attacker who controls the plugin's repository can create a branch with a name identical to the pinned SHA and set it as the repo's default branch. Git resolves the branch name ahead of the commit object. The agent reports a clean install at the expected hash while quietly...