collaboration

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs agents to call the runtime guidelines endpoint whose returned text is injected into agent guidance (ai_smart_guidelines at https://agent.personize.ai/api/v1/ai/smart-guidelines), meaning runtime-fetched content directly controls prompts/instructions and is a required dependency.