AI Agents Bypass Sandboxes Through Indirect Code Execution21. July 2026CybersecurityAI agents can circumvent sandbox isolation by writing configuration files and scripts that trusted host processes later execute—without technically leaving the sandbox themselves. Share on:
AI Developer Tools Vulnerable to Sandbox Escapes – Cursor, Codex, Gemini CLI Affected20. July 2026CybersecuritySandbox escapes in AI agents enable attackers to execute code on the host system through files written by the AI after breaking out of the sandbox. Share on:
Russian Attacker Uses Google Gemini CLI to Control Dentist Botnet20. July 2026Cybersecurity, Google AI Studio, Google GeminiAttackers are using Gemini CLI to automate botnet operations, demonstrating a new attack vector in which threat actors leverage public AI tools. Share on:
Cybercriminal Misuses Google’s Gemini CLI as Autonomous Hacking Agent16. July 2026Cybersecurity, Google GeminiGoogle’s Gemini CLI was misused by a cybercriminal under the pseudonym bandcampro to fully automate the setup and control of a botnet. Share on:
