robonet-workbench
por nickemmons
Coding Agents & IDEsResumen
Use Robonet's MCP server to build, backtest, optimize, and deploy trading strategies.
Descripción
Qué hace esta skill
Robonet's MCP server exposes 24 tools for building, backtesting, and deploying crypto and prediction market trading strategies. Tools span data exploration, AI-powered strategy generation, historical backtesting, Polymarket YES/NO trading, and live deployment on Hyperliquid. Strategies are written in Python using the Jesse framework and can be augmented with Allora Network ML price predictions.
Combines AI strategy generation, parameter optimization, ML signal integration, and live deployment in one tool chain without requiring manual exchange API setup or custom backtesting infrastructure.
Cuándo usarla
- Building a custom RSI momentum strategy for BTC-USDT on Hyperliquid
- Backtesting a mean-reversion strategy across 6 months of ETH price history
- Adding Allora ML price predictions to an existing trend-following strategy
- Trading Polymarket prediction markets on political or economic outcomes
- Stopping or restarting a live trading deployment without touching exchange APIs
Flujo de trabajo de ejemplo
User asks: Create and backtest a momentum strategy for ETH-USDT on the 4h timeframe
- 1
Call get_all_symbols to confirm ETH-USDT is available on Hyperliquid
- 2
Call get_all_technical_indicators to identify relevant momentum indicators
- 3
Call create_strategy with a description of the desired entry and exit logic
- 4
Call run_backtest on 6 months of historical ETH-USDT 4h data
- 5
Review returned metrics: Sharpe ratio, max drawdown, win rate, and total return
Python strategy code stored in Robonet with backtest results showing performance metrics across the test period
Requisitos
- Robonet API key from robonet.finance dashboard
- Robonet USDC credits purchased via dashboard billing
- Hyperliquid wallet with 200+ USDC for Vault-based deployments
Instalación
clawhub install nickemmons/robonet-workbenchEjecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.