Building a multi-agent system (MAS) for the Indian stock market (NSE/BSE) requires a "Coordinator" architecture where specialized agents collaborate to filter the 5,000+ listed stocks into a high-conviction intraday watchlist.Below is the Product Note for your AI Trading Recommendation Agent.Product Note: BharatQuant MAS (Multi-Agent System)Concept: A collaborative AI ecosystem designed to identify high-probability intraday breakouts and breakdowns by synthesizing technical, fundamental, and sentiment data in real-time.1. Core Architecture: The "War Room" ModelThe system operates as a digital trading floor where specialized agents debate the merits of a stock before recommending it.The Agent Lineup:The Scout (Market Data Agent): Constantly scans the NSE/BSE for volume spikes, price surges, and "Open = Low/High" patterns. It feeds the initial "interesting" list to the rest of the team.The Chartist (Technical Analyst): Specialized in Indian market nuances (e.g., VWAP, Supertrend, and Pivot Points). It identifies "Breakup" (resistance breach) and "Breakdown" (support failure) setups.The Globalist (Macro Agent): Monitors GIFT Nifty, US Futures (Nasdaq/S&P 500), and Brent Crude prices. It provides the "Market Mood"—if the Nasdaq is down 2%, it may flag a "Caution" signal on Indian IT stocks.The Librarian (Memory & Learning Agent): Maintains both sucesses and failures and "Mistake Log." If the system previously recommended a breakout that failed due to low delivery volume, this agent vetoes similar setups today.The Newsroom (Sentiment Agent): Scrapes real-time feeds (Moneycontrol, ET, Twitter/X) for block deals, earnings surprises, or regulatory (SEBI) updates.The Judge (Orchestrator): The final decision-maker. It gathers votes from all agents and issues the final "Strong Buy/Sell" recommendation with a confidence score.2. Key Features & FunctionalityFeatureDescriptionReal-time ScanningIdentifies stocks breaking out of 15-min or 30-min opening ranges.Sentiment CorrelationCross-references a price breakout with positive news sentiment to avoid "fakeouts."Global SyncAdjusts recommendations based on the