What problem does it solve? Answering complex questions thoroughly requires more than a quick search: sources must be gathered, read in full, cross-checked against independent references, and synthesized into a report with citations. This Skill provides a disciplined multi-pass research workflow that prevents shallow, single-source answers. ## Core Features & Use Cases - Structured multi-pass research: Declares a research mission, locks the question to one sentence, then runs broad and narrow search passes over web and academic sources like arXiv. - Adversarial cross-checking: Every load-bearing claim is verified against a second independent source or marked as unverified, with disagreements reported openly. - Cited report generation: Produces a markdown report with answer, key findings, disagreements, and a source table, saved to a file and summarized into memory. - Use Case: Ask for a deep dive on a technical topic such as world models in AI; the Skill searches arXiv and the web, reads the actual papers, cross-checks claims, and delivers a dated, cited report file. ## Quick Start Research the current state of world models in AI thoroughly and write me a cited report saved to a file.