Tools Reference

Tools Reference

11 tools + 1 prompt. All analysis defaults to the Indian market (NSE .NS / BSE .BO); US tickers also work.

Symbol formats

Market	Format	Example
NSE (India)	<SYMBOL>.NS	RELIANCE.NS, TCS.NS
BSE (India)	<CODE>.BO	500325.BO
US	plain ticker	AAPL

Analysis tools

get_quote(symbol)

Current price, previous close, day change %.

technical_analysis(symbol, period="6mo")

SMA 20/50/200, RSI-14, MACD + signal + trend, 52-week high/low, daily volatility.

analyze_stock(symbol) — the core tool

Technical + fundamental scoring → 0–100 score, BUY (≥60) / HOLD / SELL (≤20), plain-English reasons, fundamentals block, and a data_snapshot (every input value, timestamped, with source — the audit trail).

analyze_watchlist(symbols="")

Ranks up to 15 comma-separated symbols by score. Empty → default NSE large caps (RELIANCE, TCS, INFY, HDFCBANK, ITC, SBIN).

compare_stocks(symbols)

Side-by-side P/E, market cap, profit margin, ROE, debt/equity, dividend yield, 1-year return.

market_movers(market="india")

Index snapshot — India (NIFTY 50) or US (S&P 500), with suggested next steps.

stock_news(symbol, limit=5)

Recent headlines (title/publisher/link). Headlines only — tone judgment is left to the client LLM.

alpha_vantage_overview(symbol)

Optional Alpha Vantage company overview. Needs ALPHA_VANTAGE_API_KEY. US symbols only (see APIs-and-Data-Sources).

Decision journal

log_decision(symbol, action, rationale, score)

Records a BUY/HOLD/SELL decision with the price at decision time. Stored in MongoDB (stock_data.decision_journal) or decision_journal.json if MongoDB is unreachable.

review_decisions(symbol="")

Prices every open decision, computes return since decision, and verdicts: BUY → CORRECT (>+1%) / WRONG (<-1%) / NEUTRAL; SELL → inverse; HOLD → NEUTRAL while within ±5%, BROKEN beyond. See Decision-Journal.

Workflow

analyst_reports(symbol)

Three independent, timestamped reports: fundamentals, technical, sentiment (headlines). The input for the debate prompt.

Prompt: bull_bear_debate(symbol)

TradingAgents-inspired workflow the client LLM executes: gather reports → bull case → bear case → risk check (falsification conditions, sizing, upcoming events) → decision with score + confidence + horizon → log_decision automatically. Grounding rule: every claim must trace to the snapshot or a fresh tool call.

Scoring model (current, beta)

Starts neutral at 50; adjustments: price vs 50-day SMA ±10, vs 200-day SMA ±10, RSI <30 +10 / >70 −10, MACD crossover ±5, positive 6-month return +5, P/E <25 +10 / ≥60 −5, debt-to-equity <100 +5 else −5, profit margin >10% +5, dividend yield +5.

⚠️ See Backtesting: these weights are conventions, not yet fully evidence-tuned.