
Now that attackers are leveraging AI, automation is essential for defenders. How will the use of AI change security operations, such as anomaly detection, alert prioritization, and automatic response through SOAR? We will explain realistic defense strategies in which humans and AI work together. |Colorkrew Security

Explains the vulnerability MCPosion (CVE-2025-54136) discovered in the AI code editor "Cursor". We have identified flaws in the trust model and summarized the methods and countermeasures that can lead to remote code execution (RCE) by exploiting MCP settings. We are now publishing methods to protect ourselves from new threats lurking in AI-assisted development environments. |Colorkrew Security

Explains the threat of new malware PROMPTFLUX and PROMPTSTEAL, which exploit generative AI such as Gemini to autonomously mutate and generate instructions. Organize attack methods and practical countermeasures based on GTIG reports|Colorkrew Security