Where thirteen artificial minds debate, dissent, and deliver consensus on the world's most volatile energy market.
Imagine a courtroom of silicon judges, each with distinct temperaments, training, and temperaments—now picture them arguing over the direction of West Texas Intermediate crude oil futures. This is not a thought experiment. This is The Sentient Quorum, a thirteen-agent adversarial AI committee engineered to produce the most rigorously debated, stress-tested, and contrarian-sourced trading theses for discretionary WTI crude positions.
Built upon the architectural bones of Claude Sonnet's precision reasoning, Grok 4.20's edge-case exploration, and an Opus-level judge that arbitrates between them, the system operates with a heartbeat-style position manager that pulses through market data, scalp-brain micro-strategies, and forward-projecting theses. A full-context dashboard surfaces every debate, every objection, every suppressed alternative hypothesis—so that the human trader never executes in darkness.
This is not automation. This is persuasion engines at scale, producing conviction through conflict.
| Agent Role | Personality | Function |
|---|---|---|
| Claude Sonnet | Analytical maximizer | Builds base-case theses from fundamental data, supply chains, and macro indicators |
| Grok 4.20 | Contrarian provocateur | Generates tail-risk scenarios, black swan alternatives, and suppressed narratives |
| Opus Judge | Impartial arbiter | Evaluates debate quality, scores argument coherence, and assigns conviction weight |
| Scalp Brain | Tactical optimizer | Monitors intraday micro-patterns for entry/exit enhancement |
| Forward Thesis Engine | Horizon scanner | Extrapolates current debates into 30-90 day projection maps |
| Eight Supporting Agents | Specialized critics | Each attacks a specific dimension (geopolitical, seasonal, storage, freight, refinery margins, dollar correlation, weather, algorithmic flow) |
Each agent is compelled to defeat the majority opinion at least once per session. Consensus is not harmony—it is the scar tissue of arguments that survived.
Traditional risk management is a rearview mirror. The Opus Heartbeat operates differently:
- Continuous pulse checks every 90 seconds against live market data
- Position sizing recalibration based on real-time debate confidence scores
- Threshold alerts when any agent dissents above an 80% confidence interval
- Automatic thesis invalidation if the judge determines a core assumption has broken
The heartbeat is not a stop-loss. It is a cognitive vital sign monitor—if the committee stops agreeing on reality, the position shrinks until consensus reforms.
The dashboard does not show you prices. It shows you why prices exist:
- Debate Logs: The complete transcript of every agent argument, rebuttal, and concession
- Conviction Thermometer: Visual heatmap of which arguments are winning or losing over time
- Thesis Tree: A branching map of assumptions and their dependencies, with nodes marked as "verified," "challenged," or "broken"
- Scalp Brain Overlay: Real-time micro-pattern detection mapped against committee sentiment
- Forward Projection Map: 30/60/90-day scenario probabilities derived from debate outcomes
No black boxes. Every recommendation is traceable to a specific exchange between named agents.
- Adversarial by design: Agreement is suspicious. The system actively cultivates dissent
- Human-in-the-loop: The trader is the final decision node, not a bystander
- No latency arbitrage: This is about conviction, not speed
- Full transparency: Every agent's reasoning is exposed, including failures and retractions
- Multilingual debate capability: Agents can argue in English, Mandarin, Arabic, or Spanish—with cross-verification
- 24/7 operational cadence: Markets sleep. This committee does not
- 🔄 Multi-agent adversarial committee with role specialization
- 📡 Real-time heartbeat position monitoring at 90-second intervals
- 📈 Scalp brain micro-strategy integration for precision execution
- 🗺️ Forward thesis projection engine with scenario probability maps
- 📋 Full-context dashboard with traceable reasoning logs
- 🌐 Multilingual debate support across four major languages
- 🧩 Assumption dependency tree with live verification status
- ⚖️ Opus-level judgment arbitration with conviction scoring
- Discretionary WTI traders who want AI-assisted conviction without surrendering control
- Energy analysts who need adversarial stress-testing of their fundamental theses
- Risk officers who want to see the full assumption chain before approving size
- Prop trading desks looking to supplement intuition with structured debate
- Commodity research teams exploring alternative forecasting methodologies
This project is released under the MIT License. You are free to use, modify, and distribute this system, provided you retain the original copyright notice. The committee architecture, agent personalities, and heartbeat methodology are open for adaptation to other asset classes.
License: MIT License
© 2026 The Sentient Quorum Project. All rights reserved.
Trading in WTI crude oil futures and related instruments carries substantial risk of loss and is not suitable for all investors. This system is a decision-support tool, not an automated trading system. It does not execute trades, manage capital, or guarantee performance. The adversarial committee may produce incorrect, incomplete, or contradictory conclusions. Past debate performance does not predict future market outcomes. You assume full responsibility for any trading decisions made with or without reference to this system. The developers, contributors, and licensors are not liable for losses incurred through use of this software. Always consult a qualified financial advisor before engaging in speculative trading activity.