creating-internal-agents

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The codebase intentionally exposes a MedusaExec facility that writes and require()s arbitrary TypeScript and runs it with full Medusa DI container access — a powerful RCE primitive that, if misconfigured or reachable by an attacker or a malicious agent, enables data exfiltration, credential theft, and full system compromise despite the documented admin-only safeguards.