What problem does it solve?
Planning thorough research on complex questions is hard to structure: it is unclear how to decompose the question, how many researchers are needed, and how to coordinate their work. This Skill turns any research question into a complete, executable multi-agent research plan without running the research itself.
Core Features & Use Cases
- Question Decomposition: Parses the research question, classifies its domain (codebase, technical, market, user research, architecture), assesses complexity, and breaks it into 3-7 independent sub-questions with a dependency graph.
- Dynamic Team Design: Composes a tailored team of researcher agents with self-contained spawn prompts, model selection (sonnet vs opus), output formats, and measurable completion criteria.
- Executable Task Plan: Produces a wave-based task list (parallel research, analysis, synthesis) with acceptance criteria, cross-cutting citation and confidence standards, and an orchestration guide for Claude Code agent teams.
- Use Case: Ask "Compare state management libraries for React" and receive a structured research plan file with assigned researchers, parallel tasks, and a synthesis step ready for agent team execution.
Quick Start
Ask the assistant to plan a multi-agent research effort on your question, for example: run /prp-research-team with the question "What are the best approaches for real-time collaboration?"
