What problem does it solve? Answering complex questions from memory alone produces ungrounded, unverifiable claims. This Skill runs a structured deep-research protocol that decomposes a question, gathers evidence via web search and fetch, verifies every claim against its source, and delivers a cited report. ## Core Features & Use Cases - Iterative Retrieval Loop: Decomposes the question into subtopics and runs up to 4 rounds of targeted web_search and web_fetch, maintaining an evidence ledger keyed by subtopic. - Parallel Synthesis & Verification: Dispatches parallel sme workers to synthesize evidence-grounded findings, then runs dual-reviewer claim verification and a critic challenge for high-stakes claims. - Honest Reporting: Leads with the answer, cites every load-bearing claim, surfaces source disagreements, and explicitly marks unverifiable subtopics as UNVERIFIED. - Use Case: Ask "Compare current approaches to post-quantum TLS deployment" and receive a structured report where each claim traces to a fetched primary source, with conflicts between sources presented openly. ## Quick Start Ask the agent to run deep research on your question, for example: "Run deep research on the current state of WiSARD versus Tsetlin machine accuracy benchmarks."
