.svg){kind=logo}
Articles by
Lakera Guard — Fall ’25: Adaptive at Scale
Protect every AI agent - now adaptive to your apps, fluent in 100+ languages, and engineered for speed.
Agentic AI Threats: Memory Poisoning & Long-Horizon Goal Hijacks (Part 1)
How memory poisoning and long-horizon goal hijacks expose the hidden weaknesses of agentic AI, and what Lakera’s research reveals about defending them.
Cursor Vulnerability (CVE-2025-59944): How a Case-Sensitivity Bug Exposed the Risks of Agentic Developer Tools
Lakera researcher Brett Gustafson uncovered CVE-2025-59944 in Cursor, revealing how a simple case-sensitivity bug can expose deeper risks in agentic developer tools.
Inside Agent Breaker: Building a Real-World GenAI Security Playground
A behind-the-scenes look at how Lakera built Agent Breaker into a real-world GenAI security playground with layered defenses and scoring metrics.
Zero-Click Remote Code Execution: Exploiting MCP & Agentic IDEs
A zero-click exploit shows how a single Google Doc can trigger MCP abuse in Cursor, leading to stolen credentials, persistence, and enterprise risk.
Gandalf: Agent Breaker—Think Like a Hacker, Prompt Like a Pro
Play Gandalf: Agent Breaker, Lakera’s hacking simulator that lets you exploit real-world AI agent flaws, learn GenAI security, and climb the leaderboard in this immersive game.
Activate
untouchable mode.
untouchable mode.
Get started for free.
Lakera Guard protects your LLM applications from cybersecurity risks with a single line of code. Get started in minutes. Become stronger every day.
Join our Slack Community.
Several people are typing about AI/ML security. Come join us and 1000+ others in a chat that’s thoroughly SFW.
