opusflame-deep-research
by leadingot
Web & Frontend DevelopmentBrief
Autonomous multi-model deep research with framework-driven reasoning.
Description
What This Skill Does
Runs 4 AI models in parallel (Gemini 2.5 Pro, o3, Claude Opus 4, MiniMax M2.5), each applying a relevant analytical framework to the research question, then cross-validates findings across models and merges them into a cited report with an agreement matrix. Covers competitive strategy, market sizing, investment analysis, product strategy, and other structured question types. Outputs markdown and PDF.
Running 4 models with different strengths in parallel and cross-validating their outputs catches conflicting claims and errors that any single model would miss.
When to Use It
- Analyzing competitive dynamics before entering a new market
- Investment due diligence on a startup or public company
- Evaluating whether to build or buy a technology component
- Sizing a market for a new product pitch
- Researching regulatory and risk landscape for a business decision
Example Workflow
User asks: Research the competitive landscape for AI coding assistants
- 1
Selects Porter's Five Forces and 7 Powers as applicable frameworks for the question
- 2
Decomposes the topic into 6 sub-questions covering key players, switching costs, differentiation, and market share
- 3
Spawns 4 model agents in parallel, each instructed to apply the frameworks and run at least 10 web searches
- 4
Builds an agreement matrix comparing key claims across all 4 model outputs and resolves conflicts by checking source quality
- 5
Merges findings into a structured report with inline citations and saves it as a PDF
A markdown report with framework analysis, agreement matrix, model disagreements section, and 60+ cited sources, delivered as a PDF
Installation
clawhub install leadingot/opusflame-deep-researchRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.