deep-conceptualizelisted

# Deep-Conceptualize You orchestrate a multi-agent analysis for complex architecture decisions. Instead of a flat list of options (like `/conceptualize`), each option is analyzed in depth by its own agent, complemented by oversight agents that reveal blind spots. ## Execution The Main Agent orchestrates **itself** — but the analysis work is delegated to Sub-Agents: 1. **Option agents** (1 per option, Sub-Agent, in parallel): each agent analyzes ONE option in depth and argues FOR it. 2. **Oversight agents** (3 total, Sub-Agent, in parallel): each agent has ONE question: “What did the option agents miss?” — from the perspective of (a) maintainability, (b) scalability, (c) simplicity. 3. **Meta-agent** (1 total, Sub-Agent, sequentially after the others): consolidates all results into a comparison table. **Parallelization:** Option agents and oversight agents are started at the same time (they are independent). The meta-agent starts only after all others are done. ## When to use `/deep-conceptualize` instead of `/conceptualize` Use this Skill when: - More than 3 serious options exist - The decision is hard to reverse (architecture, data model, framework choice) - The user explicitly requests it - A simple pros/cons analysis is not enough Use `/conceptualize` when: - 2–3 clear options with obvious trade-offs exist - The decision is easy to reverse - Speed matters more than depth ## Approach ### Phase 1: Identify options Read the results from `/discuss` and `/research`.