debug

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). This skill injects fetch() calls that capture and transmit runtime state (including user IDs, params, request/response bodies and errors) to a logging endpoint and even instructs checks to avoid duplicate localhost logging—functionality that can be easily repurposed to deliberately exfiltrate credentials/tokens or persist a covert backdoor if instrumentation is left in or the endpoint is pointed to an attacker-controlled server.