🧠 Nexus Command Lab: Tactical Operations Center

"The difference between victory and defeat is often a single correct decision in the draft."

Nexus Command Lab is an advanced tactical intelligence suite designed for Heroes of the Storm. It transforms raw match telemetry into actionable psychological and tactical advantages through multi-agent AI analysis and forensic match review.

🚀 Key Features

⚔️ Command Center

Real-time drafting intelligence. Upload a screenshot of your loading screen or draft lobby to:

• Identify Combatants: Scans teammates and enemies for known historical data.
• Predict Outcomes: Calculates win probabilities based on map-specific mastery and team compositions.
• Social Intelligence: Flags "Synergy Anchors" (reliable allies) and "Apex Threats" (recurring nemesis rivals).

🎒 The Arsenal

A categorized view of your hero pool, preventing "experimental picks" on maps where you should be playing your best:

• Proven: High games, high win rate. Your reliable anchors.
• Hot Streaks: Momentum-based picks from the current season.
• Pockets: Low sample size but high performance "secret weapons".
• Untapped: High general skill but latent map-specific potential.
• Avoid: Statistically confirmed weaknesses.

🧭 Map Directives

For every map in the rotation, Nexus Command Lab defines a specific Win Condition (e.g., RACE, ZONE, PUSH) and recommends heroes that excel at executing that specific strategy.

🤖 Multi-Agent AI System

Nexus Command Lab uses a specialized swarm of AI agents, each optimized for specific domains:

• FORENSIC ANALYST: Post-game forensic analysis, identifying subtle mechanical failures and critical turning points.
• SCOUT: Draft intelligence, hero recommendations, ban priorities.
• TACTICIAN: Map-specific strategies, rotation timing, objective prioritization.
• SOCIAL: Player synergy analysis, nemesis tracking, duo queue recommendations.
• AUDIT: Quality assurance agent that verifies AI predictions against replay truth.
• MECHANICAL: Specialized micro-analysis for specific heroes (Stitches Hooks, Kharazim Dashes, Azmodan Stacks).

📊 Data Provenance

Multi-source verification ensures your stats are accurate:

• Blizzard Source: Official totals and rank verification via screenshot ingestion.
• Forensic Replay Parser: Granular talent-level data, DC detection, and teammate/enemy interaction history (v3.2).
• Audit Logs: Transparent history of how stats were calculated.
• Source Attribution: Every stat displays its source (Verified/Parsed/Baseline).

🧠 Cognitive Architecture

Nexus Command Lab leverages a best-in-class multi-model strategy, routing tasks to the most effective intelligence:

• Multimodal Vision: Identifies allies and enemies from screenshots.
• Structured Relational RAG: Ingests massive JSON telemetry streams into large token windows for forensic-level analysis.
• Autonomous Swarm Intelligence: Multi-agent system (Scout, Analyst, Tactician, etc.) for specialized recommendations.
• Truth Validation: Dedicated "Auditor" agent that cross-references AI narrative against hard replay data.

🛠️ Technology Stack

• Frontend: React, Vite, Tailwind CSS, Framer Motion
• Backend: Python (Flask), SQLite (Nexus Core)
• AI: Multi-model orchestration (Gemini 2.0, GPT-4o, Claude 3.5)
• Processing: Custom Python Replay Parser (MPQ based)

🚦 Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• Google Gemini API key
• Heroes of the Storm (for replay file generation)

Installation

1. Clone the repository:

   git clone https://github.com/jmoncayo-pursuit/Nexus-Command-Lab.git
   cd Nexus-Command-Lab

2. Run the unifying dev script:

   chmod +x dev.sh
   ./dev.sh

3. Configure your .env file with your Gemini API key:

   GOOGLE_API_KEY=your_key_here
   PLAYER_NAME=YourInGameName

🏗️ Architecture: DX + AX

Nexus Lab Experiments prioritize DX (Developer Experience) and AX (Agent Experience) as foundational architecture standards for a robust and user-friendly environment, influencing system design and implementation.

DX — Developer Experience

• Unified entrypoint: ./dev.sh brings up the frontend, API, and replay watcher.
• Predictable API: REST under /api with a single discovery manifest.
• Structured Relational RAG: SQLite for deterministic retrieval instead of vector databases.
• Self-Healing: Automated data integrity monitoring and repair (Nexus Healer).

AX — Agent Experience

• Dual-interface: Visual layer (React) for humans; structured data layer (JSON API + tools) for agents.
• llms.txt: LLM-oriented site guide at /llms.txt.
• API discovery: GET /api/discovery returns machine-readable endpoint descriptions.
• WebMCP tools: The app registers callable tools (e.g. nexus_match_history, nexus_ask, nexus_map_stats) via the Web Model Context API.

---

Nexus Command Lab: Evolve your game.