모델 성능 향상이 곧 취약점 확대인 AI Agent 보안의 역설과 구조적 방어 체계
AI Agents Today Aren't Secure. They're Just Clumsy
AI Agents Today Aren't Secure. They're Just Clumsy
I shipped 35 bugs in my AI chatbot. The scariest one was on the output side.
What the ChatGPT for Sheets data-exfiltration bug teaches about AI security
Microsoft Copilot just exfiltrated a company's files. The attack was one email. Here's the mechanism.
AgentThreatBench: The First OWASP Agentic Top 10 Security Benchmark
We scanned 50+ MCP servers and found HIGH-severity bugs in Atlassian, GitHub, Cloudflare, and Microsoft — here's what we learned
Prompt injection through website content: how AI agents can be manipulated by the pages they visit
Your AI Assistant is Gullible: Building a "Semantic Airgap" for Gmail Connectors
Ramp's Sheets AI Exfiltrates Financials
I tried to hack my local AI agent with Prompt Injection. It laughed at me.
Comment and Control: The GitHub AI Agent Attack That Three Vendors Hushed