What problem does it solve? Turning a vague research topic into a rigorous, actionable plan is hard: sub-questions overlap, scope drifts, and downstream search or writing stages lack a clear contract. This Skill converts a raw topic into a structured JSON research plan that later workflow stages can execute against without asking the user anything further. ## Core Features & Use Cases - Structured Plan Generation: Produces a thesis question, 3–10 MECE sub-questions, coverage dimensions, out-of-scope exclusions, and success criteria as a single JSON object. - Depth and Audience Dials: Adjusts sub-question count and phrasing based on quick/standard/deep depth and general/technical/executive audiences. - Failure Handling: Emits an explicit plan_skipped failure shape when the topic is empty or unintelligible, so downstream stages stop cleanly. - Use Case: As stage 1 of a multi-stage researcher workflow, feed it "Vector databases for RAG" at standard depth and receive a five-question plan that search, extraction, synthesis, and review stages consume directly. ## Quick Start Ask the agent to plan research on your chosen topic, optionally specifying a depth level and target audience.
