What problem does it solve? Complex architectural decisions, performance bottlenecks, and system design questions often get shallow, first-answer treatment. This Skill enforces a structured deep-thinking protocol so the agent decomposes problems, generates multiple hypotheses, and critically evaluates them before acting. ## Core Features & Use Cases - Phased Deep Analysis: A three-phase workflow covering problem decomposition, multi-hypothesis generation with mental simulation, and rubric-based decision making. - Calibrated Depth Levels: Four depth tiers (Light, Medium, Deep, Ultra) match analysis effort to problem complexity, with explicit guidance on when NOT to use deep thinking. - Structured Reasoning Artifacts: Provides a Knowledge Map template (knowns, unknowns, assumptions) and an XML-based AoT tag structure (thinking, plan, reflection) for auditable reasoning. - Use Case: When deciding between a monolith and microservices architecture, the agent maps dependencies, generates three candidate approaches, stress-tests each against edge cases, and selects one using performance, risk, and sustainability criteria. ## Quick Start Analyze whether we should migrate our monolithic app to microservices using the deep analysis protocol with hypothesis generation and risk evaluation.
