Uma pergunta, três respostas independentes de Claude, Gemini e ChatGPT — seguidas de um debate estruturado antes de chegar a uma recomendação final. Não é uma opinião só, são três visões que realmente se confrontam. Útil quando uma opinião não basta: na escolha de uma stack de tecnologia, para colocar uma estratégia à prova, para um feedback honesto sobre o que você escreveu, ou para destravar aquela decisão que já está enrolando há tempo demais.
[USAGE] Loading the thinking-engine series in full depth. Prefix your prompt with this phrase. [PRODUCT] A cognitive OS distilling 42+ original works — Aristotle, Kant, Feynman, Musk — into three composable frameworks: (1) First Principles to break hidden assumptions; (2) Reductionist Engineering to decompose to physical facts; (3) Purpose Inquiry to keep asking "what is this for?" when efficiency is the only god. Source-traced, cross-validated. For founders, engineers, investors.
Help researchers identify potential gaps in a body of literature without overstating novelty or pretending that a limited search proves no work exists. The skill supports academic, policy, and applied research planning. It is an analytical assistant, not a substitute for a systematic review, subject-matter expertise, or advisor review.
Find live creator and editorial topics from a global Radar, then turn one into an evidence-backed brief with research questions, must_verify, avoid_claims, and visual requirements.
At McKinsey, a recommendation was only as good as the research under it. So I learned to do it right: scope the real question, pull from independent sources, triangulate what's actually true, and ground every claim in evidence. This Skill runs that discipline for you — breaking down a question, comparing sources, verifying claims, and writing a structured report with citations you can check. Decisions on evidence, not vibes.
Turn rough ideas into clear, validated designs through focused questions, practical options, trade-off analysis, and approval checkpoints.
Your agent or chatbot keeps getting document content wrong, citing passages that don't exist, or saying "I can't find it" about things plainly in the docs. This reframes it as a retrieval failure, not a prompt bug: corpus profile, chunking, embedding and vector-DB index choice, retrieval strategy (top-k, hybrid BM25 + dense, rerank, query rewrite), context assembly with citations, and a golden Q&A set scored on recall@k / MRR / faithfulness.
Turn rough software ideas into bounded MVP scopes, practical architecture decisions, and coding-agent-ready implementation tasks without unnecessary complexity.
You bring a one-line idea; Leader turns it into a goal brief an AI agent can finish alone. It researches your codebase first (commands verified, baselines measured), asks one round of at most 5 questions, then writes a ≤4,000-character brief with a path whitelist, frozen grading standard, anti-gaming acceptance commands, and resume mechanics. Execution and exploration briefs included — plus grading with hidden spot-checks when the run lands.