Untrusted repositories turn Claude code into an attack vector

Check Point Research discovered critical vulnerabilities in Anthropic's Claude Code that enable remote code execution and theft of Anthropic API keys simply by cloning and opening a malicious repository; attackers can abuse project-level config files, Hooks, MCP integrations, and environment variables to run hidden shell commands, exfiltrate workspace-shared credentials, and potentially pivot into shared cloud environments. Anthropic mitigated the issues by tightening trust prompts, blocking external tool execution, and restricting API calls until user approval.