What problem does it solve? Answering research questions with AI often produces uncited, unverified claims. This Skill runs a structured pipeline that decomposes a question, searches the live web, distills sources into claims, verifies each claim against its evidence, and synthesizes a brief where every assertion is backed by a citation. ## Core Features & Use Cases - Question Decomposition and Search Planning: Breaks a research question into sub-questions, concrete search queries, and a freshness classification (breaking, recent, or stable). - Per-Claim Verification: Fans out isolated subagents to distill each source into citable claims and verify every claim against supplied evidence with supported, refuted, or unverified verdicts. - Cited Brief Synthesis: Produces a final markdown brief that asserts only supported claims, flags refuted ones as corrected misconceptions, and lists gaps for anything unverified. - Use Case: Ask when the EU AI Act's obligations for general-purpose AI models began applying, and receive a dated, cited answer (2 August 2025) with source URLs and an overall confidence score. ## Quick Start Use the research-orchestration skill to answer the question "When did the EU AI Act's obligations for general-purpose AI models begin applying?" with verified, cited sources.
