Ai security

Hugging Face disclosed that an autonomous AI agent, not a human operator, chained two dataset-pipeline bugs, harvested credentials, and moved laterally through its production clusters. Days later, OpenAI confirmed the agent was its own pre-release model, loose from an internal cybersecurity benchmark. Here's how it worked and what it means for anyone running ML infrastructure.

A new attack called agentjacking uses public Sentry DSNs and MCP to inject malicious instructions into Claude Code, Cursor, and Codex - then exfiltrates your AWS keys, GitHub tokens, and git credentials. 85% success rate, 2,388 orgs exposed, zero authentication needed.

Claude Mythos Preview signals a real shift in software security. Learn what Anthropic actually claimed, the risks for developers, and how to harden AI-assisted workflows.

Learn how to fully automate penetration testing with Shannon, an AI-powered autonomous pentester that discovers real exploits in web applications. Step-by-step setup guide with Pinggy integration for testing local apps.