Indirect Prompt Injection in AI Coding Assistants via Malicious Documentation Leads to IDE Command Execution | AI Breaking Wire