What problem does it solve? Answering complex research questions with a single search often yields shallow or incomplete results. This Skill decomposes a big question into a dependency graph of sub-questions, researches them in parallel, and synthesizes a cited report. ## Core Features & Use Cases - DAG-Based Planning: Breaks a research question into 4-12 sub-questions with explicit dependencies, printed as a plan table before execution. - Parallel Subagent Research: Runs independent sub-questions concurrently as web-search subagents, each returning findings, gaps, and cited sources. - Gap-Driven Iteration: Reviews collected gaps and runs one targeted follow-up wave when missing information would change the conclusion. - Persistent Reports: Saves the synthesized markdown report with deduplicated sources into the project's docs/deep-dive/ directory. - Use Case: Ask "How do microservices generate Jira documentation?" and receive a structured report with an executive summary, themed findings, open questions, and a full source list. ## Quick Start Ask Claude to run a deep dive on the research question or topic you want investigated thoroughly.
