Convene a council of 5 AI advisors to pressure-test any decision and eliminate Claude's yes-man bias. Five advisors with opposing thinking styles analyze in parallel, 5 peer-reviewers evaluate anonymized outputs, then a Chairman synthesizes one verdict and one concrete next step. Based on the LLM Council methodology by Ole Lehmann and Andrej Karpathy. Trigger on: "council this", "run the council", "war room this", "pressure-test this", "stress-test this", "debate this", "should I X or Y" (when a real tradeoff exists).