council

To validate architectural decisions with multi-agent review, provide your architecture specs to trigger a five-expert council analysis. The process loads context, executes parallel analysis, cross-critiques the findings, and synthesizes a structured YAML output containing a clear, actionable decision with rationale.

A multi-expert council for software architecture is a coordinated review mechanism where five experts analyze complex changes. It enforces a three-phase workflow—parallel analysis, cross-critique, and synthesis—to surface a documented, actionable decision that aligns stakeholders during escalations.

You should escalate architecture concerns to a multi-agent review when changes touch multiple files, teams, or subsystems. It is applicable during autopilot escalations to validate complex decisions, align stakeholders, and produce a documented rationale via structured cross-critique and synthesis.

Yes, you can use multi-agent review for cross-team architecture alignment. The five-expert council coordinates a structured cross-critique of parallel analyses, specifically surfacing a clear, actionable outcome to align stakeholders when changes span multiple subsystems or teams.

Cross-critique in an LLM architecture decision review works as the second phase of the three-phase process. After parallel analysis, the five experts critique each other's findings before synthesis, ensuring the final structured YAML output reflects a validated, actionable decision with comprehensive rationale.

The architecture council review output uses a structured YAML format. After completing the three-phase process of context loading, parallel analysis, cross-critique, and synthesis, it produces a documented architectural decision with clear rationale and actionable next steps.