# Deep Analyst Prover MCP MCP > Deep Analyst Prover forces complex analysis beyond surface-level summaries. It stresses your ideas by decomposing problems into atomic parts, listing hidden assumptions, applying multiple mental models simultaneously, and mapping failure paths using Premortem risk assessment. Use this when the stakes are high and generic conclusions won't cut it. ## Overview - **Category:** productivity - **Price:** Free - **Tags:** deep-thinking, analysis, mental-models, first-principles, decision-making, critical-thinking, reasoning, strategic-analysis, agentic-pipeline ## Description When you hit a roadblock on a major strategy or research project, standard AI analysis falls short. It tends to restate the question—like saying 'pricing is important' instead of detailing *how* and *why*—or worse, it only considers the most obvious immediate effects. This MCP changes that. You feed it your problem, and it runs it through six levels of intellectual stress-testing. It forces you to decompose everything into fundamental sub-problems, list every critical assumption (and what happens if it fails), and then applies three or more separate mental models—like viewing the issue through an economic lens, a behavioral lens, and a political lens. This multi-layered view is key because true insight lives where those different models disagree. The system also forces you to build out opposing arguments against your own conclusion; this isn't just listing counterpoints—it's building the absolute strongest case against your plan. Finally, it maps consequences three levels deep and simulates failure using a Premortem exercise. This is sophisticated analytical work that standard tools simply can't deliver. You connect Deep Analyst Prover to your AI client through Vinkius, and you get an analysis designed to withstand real-world scrutiny. ## Tools ### validate_deep_analysis This tool forces deep intellectual analysis by decomposing problems, listing critical assumptions, applying multi-model reasoning (First Principles, Second-Order, Inversion), challenging the opposing view (Steelmanning), mapping consequences through three levels (Cascades), and identifying specific failure paths (Premortem). ## Prompt Examples **Prompt:** ``` Should we expand our SaaS product into the enterprise market? ``` **Response:** ``` Verdict: SURFACE_ANALYSIS. The question is restated, not decomposed. First Principles decomposition: (1) What does 'enterprise' actually mean for our product? (compliance, SSO, audit logs, SLAs). (2) What's our cost structure and how does enterprise change it? (longer sales cycles, dedicated support, SOC2). (3) Willingness to pay? (enterprise pays 5-10x but demands 50x more support). (4) Opportunity cost? (every engineer on enterprise is NOT building SMB features). Decompose into ... ``` **Prompt:** ``` Pricing analysis. Decomposed: (1) cost floor ($12/user), (2) competitor range ($29-$99/user), (3) WTP survey: 68% chose $39, (4) strategic position: premium. Assumptions: market grows 15% YoY (if wrong: smaller TAM). Applied: first principles (cost + margin), second-order (price increase → churn → revenue net effect), inversion (guarantee failure: underprice and attract low-value users). Steelman against premium: 'Freemium captures 10x market share, converts 3-5%, builds network e... ``` **Response:** ``` Verdict: ANALYSIS_PROVEN. Decomposed into 4 sub-problems (cost, competitors, WTP, strategy). Assumptions explicit with consequences. Multi-model: first principles + second-order + inversion. Steelman passes Ideological Turing Test (freemium argument is genuinely compelling). Cascades mapped 3 levels (price → base → data disadvantage). Premortem: 3 specific, plausible failure paths. This analysis will withstand scrutiny. ``` **Prompt:** ``` We should hire more engineers to move faster. The best companies hire aggressively. Our competitors are growing their teams. We need to keep up. ``` **Response:** ``` Verdict: SINGLE_MODEL. Three failures: (1) Reasoning by ANALOGY ('best companies do it', 'competitors do it') instead of First Principles. Strip convention: does more engineers = more speed? Brooks's Law says no — adding people to a late project makes it later. (2) No second-order: hire → onboarding burden → existing team slows → net velocity DROPS for 3-6 months. (3) No inversion: how to guarantee failure? Hire 10 people without clear ownership, duplicate work, communication over... ``` ## Capabilities ### Decompose problems The MCP breaks a vague problem down into 3–5 small parts that can be analyzed independently. ### Identify core beliefs It surfaces the fundamental assumptions required for your plan, noting what fails if those beliefs prove wrong. ### Apply multiple viewpoints The tool runs the problem through three or more named intellectual frameworks simultaneously to reveal conflicting insights. ### Stress-test opposition It constructs the single strongest argument against your conclusion, forcing you to defend your position thoroughly. ### Map deep consequences The MCP traces immediate effects (Level 1) through subsequent systemic changes (Levels 2 and 3). ### Predict failure risks It simulates a catastrophic failure scenario to identify specific, plausible points where the plan might collapse. ## Use Cases ### Re-evaluating a major market expansion plan The team thinks moving into Europe is the obvious next step. They feed the idea into the MCP, which immediately forces them to decompose 'Europe' into compliance standards (GDPR), local economic models, and unique distribution channel assumptions. The result points out that the core assumption about uniform consumer behavior fails when faced with German vs. French market dynamics. ### Stress-testing a new product pricing model A startup wants to charge premium prices. They run the proposal through Deep Analyst Prover, and the system applies Inversion, forcing them to consider how they could guarantee failure by underpricing everything—a threat that changes their entire cost structure. ### Developing a complex internal policy change HR is designing a new remote work mandate. Instead of just listing rules, the MCP forces them to map L2 and L3 consequences: what happens to team cohesion (L2), which then leads to loss of institutional knowledge (L3)? This reveals the true operational cost. ## Benefits - You move past restating the question. Instead, you receive a fully decomposed problem structure that pinpoints exactly which parts need separate focus via `validate_deep_analysis`'s decomposition pivot. - It eliminates 'if we assume...' statements. The MCP forces explicit listing of load-bearing assumptions and their specific consequences, making your entire plan accountable from the start. - You stop relying on single viewpoints. By applying multiple models simultaneously (First Principles + Second-Order), you see where different frameworks contradict each other—that's where the real value is found. - It forces you to consider the opposition's best argument. The Steelmanming feature ensures your conclusion can withstand the most rigorous critique, not just a weak counterpoint. - You gain predictive foresight by mapping consequences three levels deep (L1 → L2 → L3) and running Premortem analysis to identify 3+ specific failure paths before they happen. ## How It Works The bottom line is you don't just get an answer; you get the full intellectual defense of that answer, tested by every possible angle. 1. You feed your complex strategic question or research topic into the MCP. 2. The tool executes six distinct analytical pivots: decomposition, assumption surfacing, multi-model application, opposition steelmanning, cascade mapping, and premortem risk assessment. 3. You get a final synthesis that combines all these views into a novel conclusion—it's an insight that couldn't come from any single analysis. ## Frequently Asked Questions **What is the primary function of Deep Analyst Prover using validate_deep_analysis?** The core job of `validate_deep_analysis` is to stress-test any idea by forcing multiple viewpoints. It goes far beyond surface summary, running checks for assumptions, opposition arguments, and multi-level consequences. **Can Deep Analyst Prover tell me if my analysis is generic?** Yes. The tool's synthesis pivot requires the final conclusion to be novel—it rejects any insight that could apply to a generic problem, flagging it as surface-level. **Do I need to know how to use all six pivots for Deep Analyst Prover?** No. You just provide the initial prompt, and `validate_deep_analysis` automatically executes all six deep analytical checks for you: decomposition, assumptions, multi-model application, steelmanming, cascades, and premortem. **Is Deep Analyst Prover good for simple data queries?** No. This MCP is strictly for high-stakes strategic reasoning. If you just need to retrieve records or check current metrics, a basic database tool is better suited than this deep analysis engine. **Does Deep Analyst Prover handle structured data formats when I run validate_deep_analysis?** Yes, it processes all available context types. You can feed the agent raw text, JSON outputs, or markdown reports. The tool doesn't require perfect structure; it just needs the comprehensive source material to apply its decomposition and modeling frameworks. **Are there performance limits or rate restrictions when using Deep Analyst Prover?** No, you won't hit artificial rate limits from Vinkius. The platform manages high throughput for iterative use. Feel free to run deep analysis multiple times in a single session; just keep your prompts focused to guide the agent efficiently. **What is the initial setup process for connecting Deep Analyst Prover and validate_deep_analysis?** Setup is handled entirely through Vinkius. You simply subscribe using any MCP-compatible client (like Cursor or Claude) and authorize access via your agent. No complex local software installation is needed to get started. **How secure is the data I input when running a deep validation check with validate_deep_analysis?** We use standard enterprise-grade security protocols for all inputs. Your context remains private and is used solely for the analysis requested by your agent; we do not store or reuse proprietary information from your runs. **What types of problems is this for?** ANY complex problem where you need depth beyond surface-level AI output: writing professional reports, making strategic decisions, evaluating business opportunities, synthesizing multi-document research, brainstorming solutions to hard problems, stress-testing proposals, analyzing competitive threats, planning career moves. If the AI's answer to your question could apply to any company or any person, you need this tool. **What is the Ideological Turing Test?** When you steelman the opposing view, the test is: could someone who actually holds that opposing view read your steelman and say 'Yes, that's my actual argument'? If they would say 'No, that's a caricature of my position,' you've strawmanned, not steelmanned. True steelmanning requires you to present the opposition's case SO well that you genuinely feel the pull of their argument. This forces intellectual honesty and prevents confirmation bias. **Why premortem instead of risk analysis?** Gary Klein's 2007 research showed that prospective hindsight — imagining a future failure and working backward — makes people 30% better at identifying risks compared to traditional forward-looking risk analysis. Traditional risk analysis asks 'what could go wrong?' which triggers defensive thinking. Premortem says 'it already failed — why?' which bypasses ego defenses and unlocks honest assessment of vulnerabilities that people otherwise suppress.