What problem does it solve? Choosing between options or evaluating a subject often suffers from shallow analysis, missed angles, and premature closure. This Skill runs a structured deliberation process that routes by context, discovers unseen angles, critiques and assesses each option, and delivers a committed recommendation with its reasoning trail. ## Core Features & Use Cases - Context-Routed Depth: Classifies the input by form (single, comparison, list, set, open-question) and corpus richness to select a graduated depth level from L0 to L5, so a rich single subject gets a light pass while a thin comparison gets discovery-first treatment. - Angle Discovery and Typed Hybrids: Orchestrates angle and connection discovery before assessment, then synthesizes up to three classified hybrid options (complementary, sequential, layered, reframed, constraint-splitting) as first-class candidates. - Decision Surface Output: Returns a committed recommendation, top alternatives, the assumption most likely to be wrong, and what would change the verdict, delivered as an ask-gate or inline block. - Use Case: When weighing three architecture candidates for a migration, invoke cross-examine to critique each option on shared dimensions, surface a reframed hybrid, and receive a recommendation with the key risk assumption stated. ## Quick Start Ask the agent to cross-examine your options, for example: weigh these three database candidates and recommend one with the trade-offs explained.
