claude-council

Pressure-testing high-stakes decisions involves running a structured multi-advisor analysis to surface blind spots. This process applies bias scanning, peer review, and debate to weak assumptions, preserving dissent to turn uncertainty into an actionable recommendation.

Multi-advisor debate synthesis coordinates parallel AI perspectives like Red Team, First Principles, and Outsider reviews to examine a choice from different angles. It applies bias checks and peer review to produce a synthesized recommendation with journaled outputs.

Structured decision analysis is designed for ambiguous business, product, career, and operational choices. It coordinates multiple advisor perspectives, handles disagreement through debate, and generates a journaled HTML report capturing recommendations and next steps.

Generating an HTML decision report involves running a multi-advisor fan-out with peer review and bias audits. The process automatically captures the debate transcript, confidence levels, and final recommendations into a journaled artifact for later review.

The claude-council decision-making skill requires jq as a dependency to process its structured multi-advisor analysis. This prerequisite supports the parallel advisor fan-out, peer review, and journaled output generation during high-stakes decision evaluation.

AI advisors for decision-making should not be used for choices requiring real-time data, legal compliance, or domain-specific expertise outside the provided context. The structured debate relies on the input prompt, meaning incomplete information will yield unreliable synthesis and recommendations.