llm-council

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

• Potentially malicious external URL detected (high risk: 0.90). The skill performs runtime POSTs to https://api.fireworks.ai/inference/v1/chat/completions and injects the returned model outputs into later prompts (Phase 2 rankings and Phase 3 synthesis), so external content from that URL directly controls the agent's prompts and is required for the workflow.