agentmark

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly tells the agent to WebFetch and rely on the AgentMark docs at runtime (e.g. https://docs.agentmark.co/llms.txt) to determine CLI flags, prompt syntax, and behavior, so fetched remote content from that URL directly controls agent instructions and is required for correct operation.