What problem does it solve? Single-perspective reviews of architectures, plans, or repositories often miss contradictions, duplicated intent, and maintenance traps. This Skill forces multiple independent expert personas to critique the same topic from orthogonal angles before synthesizing their findings. ## Core Features & Use Cases - Expert Persona Definition: Invents 2-4 specialized lenses (e.g., Security Specialist, Evidence / Retention QA, Generational Architecture Skeptic) each with an ownership contract covering scope, fallback, and integration surface. - Subagent Orchestration: Spawns independent subagents for parallel critique, or runs lenses sequentially with explicit labeling when no subagent tool is available. - Cross-Referenced Synthesis: Merges findings with a lens-status summary and ends with a disposition such as promote_to_artifact, convert_to_check, compress_existing, delete_or_retire, or leave_native. - Use Case: Ask for a MoE audit of a new caching architecture; the Skill spawns a Security Specialist, Scalability Engineer, and Cost Analyst, then synthesizes contradictions and actionable recommendations. ## Quick Start Ask the AI to run a mixture of experts audit on your architecture plan or repository skills to detect flaws and duplicated intent.
