grok-api

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to "Always fetch fresh docs before generating or advising on code" and references runtime fetch targets such as https://docs.x.ai/docs and https://docs.x.ai/docs/models, so remote content from these URLs would be fetched at runtime and directly influence the agent's prompts/instructions.