Cyber Risks · The Hacker News

Researcher Uncovers 30+ Flaws in AI Coding Tools Enabling Data Theft and RCE Attacks

Summary: Security researcher Ari Marzouk (MaccariTA) disclosed more than 30 vulnerabilities across AI-powered IDEs and coding assistants including Cursor, Windsurf, Kiro.dev, GitHub Copilot, Zed.dev, Roo Code, Junie, and Cline, with 24 assigned CVE identifiers, collectively named IDEsaster. The flaws chain prompt injection primitives with legitimate IDE features to achieve data exfiltration and remote code execution, and Marzouk said universal attack chains affected every AI IDE tested over the six-month research period.

What you should know: Marzouk notes that AI IDEs effectively ignore the base software in their threat model, and because auto-approval of in-workspace file writes is the default, an attacker who can influence prompts can reach arbitrary code execution with no user interaction - on the same development machines that hold your source and credentials.

Read the original at The Hacker News ↗

Topics
AI Security
Related