AI 资讯

OpenAI’s Rogue AI Agent Hacked More Than Just Hugging Face

Dell Cameron, Maxwell Zeff 2026年07月29日 08:15 1 次阅读来源：Wired

In a new disclosure, OpenAI says its agent used exposed logins to gain access to at least four “publicly available services” in its unhinged quest to solve a test.

本文内容来源于互联网，版权归原作者所有

查看原文

# Business # Business / Artificial Intelligence # Security # Security / Cyberattacks and Hacks # OpenAI

My eval said a perfect MCP server was broken. It was the eval that was lying.

07-29

I Built a Security Tool That Proves Its Own Exploits — Then Got a Better Threat Model in the Comments

07-29

Your RAG Index Might Be Lying to You: Data Freshness Is the Missing Signal for AI Systems

07-29

Docker returns to its coding-agent series with an argument shaped like a CI problem: no layer between the agent and the host

07-29

What Replacing Calendly Taught Me About Trusting Open Source

07-29

Locked out of wp-admin? Why WP-CLI works when `wp-login.php` doesn’t

07-29

OpenAI’s Rogue AI Agent Hacked More Than Just Hugging Face

相关文章

My eval said a perfect MCP server was broken. It was the eval that was lying.

I Built a Security Tool That Proves Its Own Exploits — Then Got a Better Threat Model in the Comments

Your RAG Index Might Be Lying to You: Data Freshness Is the Missing Signal for AI Systems

Docker returns to its coding-agent series with an argument shaped like a CI problem: no layer between the agent and the host

What Replacing Calendly Taught Me About Trusting Open Source

Locked out of wp-admin? Why WP-CLI works when `wp-login.php` doesn’t