xauusd-calendar-descriptions

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs using the r.jina.ai proxy (e.g., https://r.jina.ai/$url and https://r.jina.ai/http://duckduckgo.com/html/?q=...) at runtime to fetch page text that is then extracted and injected as per-event evidence — remote content that directly controls the agent's output/instructions.