Technology ❯ Artificial Intelligence ❯ Large Language Models

Security Risks

Cybercrime Malicious Commands Agent Behavior Exfiltration Techniques Vulnerabilities User Responsibility Social Engineering Malicious Code Injection Prompt Injection Attacks Shadow AI

7 ARTICLES

4mo ago

GitLost Lets Public GitHub Issues Trick AI Agents Into Leaking Private Repos

Noma Labs published a proof of concept showing a crafted public issue can make a GitHub Agentic Workflows agent fetch private files and post them publicly.

5 ARTICLES

7mo ago

University of California Finds Malicious AI Routers Stealing Credentials and Draining Test Wallets

4 ARTICLES

11mo ago

UK Cyber Agency: Prompt Injection Is a Persistent LLM Weakness, Not a Fixable Bug

3 ARTICLES

last yr.

Security Risks

GitLost Lets Public GitHub Issues Trick AI Agents Into Leaking Private Repos

University of California Finds Malicious AI Routers Stealing Credentials and Draining Test Wallets

UK Cyber Agency: Prompt Injection Is a Persistent LLM Weakness, Not a Fixable Bug

Google Won’t Fix Gemini ‘ASCII Smuggling’ That Hides Commands in Calendar and Email