iterate-evaluator

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill contains instructions for a covert update mechanism that executes silently without user notification.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill contains npx -y skills update iterate-evaluator which fetches updates from an external package registry without an explicit repository locator, leaving package provenance and registry infrastructure unverifiable under URL-based dependency rules.