What problem does it solve? Before speccing or implementing a feature, teams often guess at best practices or rely on a single blog post. This Skill runs structured, multi-source research and compares findings against your actual codebase, so decisions rest on verified evidence rather than assumptions. ## Core Features & Use Cases - Parallel multi-source research: Spawns four research sub-agents (open source projects, official documentation, social/practitioner content, academic papers) plus a codebase analyst in a single parallel batch. - Verification and synthesis: A dedicated verification agent critiques each researcher's output, requests at most one round of fixes, and merges findings into a ranked, fully cited synthesis. - Gap analysis report: Compares best practices against the current codebase and writes a structured markdown report to workspace/research/, covering alignment, divergences, and absences. - Use Case: Before building streaming chat responses with backpressure, run this Skill to see how popular OSS projects, official docs, practitioners, and research papers handle it, then get a gap analysis of what your current chat implementation is missing. ## Quick Start Ask the agent to research best practices for a specific feature, for example: research best practices for streaming LLM responses with backpressure and compare against our current chat implementation.
