- Home
- Skills
- Trading & Investing
- AI Quant Coder: ARIMA Forecasts & Paper-Trading Bots
Works with the AI Agents you already use
AI Quant Coder: ARIMA Forecasts & Paper-Trading Bots
AI Quant Coder turns your AI agent into a quantitative research assistant for price forecasting, strategy backtesting and paper-trading bot generation.
Free
AI Quant Coder: ARIMA Forecasts & Paper-Trading Bots
Example session with this skill installed
Forecast the next 30 days for prices.csv, backtest an SMA 10/25 crossover with $10,000 starting capital, and then give me a paper-trading bot script for it.
- Read your context and instructions
- Compiled the ai quant coder
Summary
Ran an AIC-optimized ARIMA forecast and an SMA crossover backtest on a 120-bar sample dataset (2023-01-02 to 2023-06-16). The model selected ARIMA(0,1,0), so the 30-day forecast is flat (random-walk behavior) with widening 95% bands. In simulation, SMA 10/25 beat buy-and-hold on this sample.
Forecast
| Metric | Value |
|---|---|
| Last close | $93.24 |
| Selected model | ARIMA(0,1,0), AIC 467.7 |
| 30-day target | $93.24 (0.0%) |
| Projected trend | NEUTRAL |
| 95% range at day 30 | $74.86 to $111.62 |
Backtest: SMA 10/25, $10,000, 0.1% fee
| Metric | Strategy | Buy & Hold |
|---|---|---|
| Total return | +5.45% | -7.78% |
| Max drawdown | -11.03% | n/a |
| Sharpe ratio | 0.73 | n/a |
| Win rate | 66.7% (2 of 3) | n/a |
| Profit factor | 1.71 | n/a |
Commands the agent ran
python scripts/forecast_arima.py --csv prices.csv --days 30
python scripts/backtest.py --csv prices.csv --strategy sma_crossover --short-window 10 --long-window 25 --initial-capital 10000
Next steps
- Only 3 trades on 120 bars is too few to trust. Re-run on 250+ bars before drawing conclusions.
- Review the generated paper-trading bot and its risk limits before changing any parameter.
- Educational simulation only, not financial advice.
Connects securely to your tools. The creator never sees your data.
About this skill
AI Quant Coder turns your AI agent into a quantitative research assistant for price forecasting, strategy backtesting and paper-trading bot generation.
What it does
- ARIMA forecasting: 30-day price projection with AIC-based order selection, 95% confidence intervals and residual diagnostics.
- Backtesting: SMA crossover, RSI and MACD strategies with transaction fees, max drawdown, Sharpe ratio, win rate and profit factor.
- Bot generation: reusable templates for modular Python trading bots that default to paper-trading simulation, with ATR stop-loss and position-sizing rules.
How it works
- Give your agent a CSV of OHLCV prices (or a ticker if yfinance is installed).
- The agent runs the bundled scripts and reads the JSON output.
- It reports the forecast, generates a bot script, and backtests it.
Good to know
- Bots are simulation-first. No live orders by default.
- Educational and research use only. Not financial advice.
- Best results with 250+ daily bars; ARIMA needs at least 50.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Trading & Investing.
Requires minimum 50 bars for ARIMA and 250 bars for backtesting. Generated bots default to simulation mode. Requires Python environment to run provided scripts.Requires numpy, pandas and statsmodels. The --ticker option needs internet and yfinance. ARIMA on price series often gives flat, random-walk-like forecasts, so treat outputs as a baseline, not a prediction.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
Trust & safety
Security scan
Security scan passed 3 days ago
- Free with an Agensi account
- Software
Python 3.9+numpypandasstatsmodels
- Optional
- yfinanceAgensi market-data tools
Runs locally, no API keys needed. Simulation only, not financial advice.
Permissions
File Scopes
Terminal runs the bundled Python scripts. Files are read (CSV prices) and written (generated bot scripts). Network is only used if you pass --ticker.