agentai

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime calls to the agent.ai actions API (POST https://api-lr.agent.ai/v1/action/), which is used to invoke remote LLMs/agents (invoke_llm/invoke_agent) and therefore executes remote instructions that control behavior.