# Deploy Live Trading > Deploy and manage live trading agents on Hyperliquid. ⚠️ HIGH RISK - REAL CAPITAL AT STAKE ⚠️ Provides deployment_create (launch agent, $0.50), deployment_list (monitor), deployment_start/stop (control), and account tools (credit management). Supports EOA (1 deployment max) and Hyperliquid Vault (200+ USDC required, unlimited deployments). CRITICAL: NEVER deploy without thorough backtesting (6+ months, Sharpe >1.0, drawdown <20%). Start small, monitor daily, define exit criteria before deploying. - Skill: `robonet-tech/deploy-live-trading` (Agent Skill) - Install (CLI): `npx skillmds@latest add robonet-tech/deploy-live-trading` - Raw SKILL.md: https://api.skillmd.com/api/skills/robonet-tech/deploy-live-trading/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: DevOps & Infra - Author: robonet-tech (https://skillmd.com/u/robonet-tech) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/robonet-tech/deploy-live-trading --- # Deploy Live Trading ``` ╔══════════════════════════════════════════════════════════╗ ║ ║ ║ ⚠️ LIVE TRADING RISKS REAL CAPITAL ⚠️ ║ ║ ║ ║ • You can lose ALL deployed capital ║ ║ • Bugs in strategy code cause significant losses ║ ║ • Market conditions change - backtest ≠ live ║ ║ • NEVER deploy without thorough backtesting ║ ║ • Start with small capital to validate live behavior ║ ║ • Monitor deployments actively (daily minimum) ║ ║ • Define exit criteria BEFORE deploying ║ ║ ║ ║ THIS IS NOT A SIMULATION ║ ║ REAL MONEY WILL BE TRADED ║ ║ ║ ╚══════════════════════════════════════════════════════════╝ ``` ## Quick Start This skill deploys strategies to live trading on Hyperliquid. Use ONLY after thorough backtesting and validation. **Load the tools first**: ``` Use MCPSearch to select: mcp__workbench__deployment_create Use MCPSearch to select: mcp__workbench__deployment_list Use MCPSearch to select: mcp__workbench__deployment_stop ``` **BEFORE deploying, complete this checklist**: - [ ] Backtested on 6+ months of data - [ ] Sharpe ratio >1.0, max drawdown <20% - [ ] Tested on multiple time periods - [ ] Code reviewed for bugs - [ ] Risk management validated (stop loss, position sizing) - [ ] Credit balance sufficient - [ ] Monitoring plan established - [ ] Exit criteria defined - [ ] Starting with small capital (<10% of intended final size) **If ANY item unchecked, DO NOT DEPLOY** **When to use this skill**: - After extensive backtesting shows consistent profitability - When ready to risk real capital - When you can monitor the deployment actively **When NOT to use this skill**: - Strategy not thoroughly tested (use `test-trading-strategies` first) - Haven't reviewed strategy code - Don't have monitoring plan - Can't check deployment daily for first week - Haven't defined when to stop deployment ## Available Tools (6) ### deployment_create **Purpose**: Deploy strategy to live trading on Hyperliquid **Parameters**: - `strategy_name` (required): Name of strategy to deploy - `symbol` (required): Trading pair (e.g., "BTC-USDT") - `timeframe` (required): Candle interval (1m, 5m, 15m, 30m, 1h, 2h, 4h, 6h, 8h, 12h, 1d) - `leverage` (optional, 1-5): Position multiplier (default: 1) - `deployment_type` (optional): "eoa" (wallet, default) or "vault" - `vault_name` (required for vault): Unique vault name - `vault_description` (optional): Vault description **Returns**: Deployment ID, status, wallet address, configuration **Pricing**: $0.50 per deployment **Constraints**: - **EOA**: Max 1 active deployment per wallet - **Vault**: Requires 200+ USDC in wallet, unlimited deployments **Use when**: All pre-deployment criteria met (see checklist) ### deployment_list **Purpose**: Monitor active deployments **Parameters**: None **Returns**: List of all deployments with status, performance, configuration **Pricing**: Free **Use when**: Checking deployment status, monitoring performance ### deployment_start **Purpose**: Resume stopped deployment **Parameters**: - `deployment_id` (required): ID of deployment to resume **Returns**: Updated deployment status **Pricing**: Free **Use when**: Restarting previously stopped deployment after validation/fixes ### deployment_stop **Purpose**: Halt live trading **Parameters**: - `deployment_id` (required): ID of deployment to stop **Returns**: Updated deployment status **Pricing**: Free **Use when**: - Live performance degrades significantly - Need to update strategy code - Market conditions change fundamentally - ANY red flag triggered (see Red Flags section) ### get_credit_balance **Purpose**: Check available USDC credits **Parameters**: None **Returns**: Current credit balance **Pricing**: Free **Use when**: Before deployment (verify sufficient credits), monitoring spending ### get_credit_transactions **Purpose**: View credit transaction history **Parameters**: None **Returns**: List of credit transactions **Pricing**: Free **Use when**: Auditing spending, tracking costs ## Core Concepts ### Deployment Types **EOA (Externally Owned Account)**: ``` Type: Direct wallet trading Setup: Immediate (no additional requirements) Limit: Max 1 active deployment per wallet Complexity: Lower Best for: Testing, personal trading, single strategy Cost: $0.50 to create Advantages: ✓ Simple setup ✓ Immediate deployment ✓ No minimum balance requirement Disadvantages: ✗ Only 1 deployment per wallet ✗ No public performance tracking ✗ Personal wallet at risk ``` **Hyperliquid Vault**: ``` Type: Professional vault setup Setup: Requires 200+ USDC in wallet Limit: Unlimited deployments Complexity: Higher Best for: Multiple strategies, professional trading, public showcasing Cost: $0.50 per deployment Advantages: ✓ Unlimited deployments ✓ Public TVL and performance tracking ✓ Professional infrastructure ✓ Separate from personal wallet Disadvantages: ✗ Requires 200+ USDC setup ✗ More complex configuration ✗ Public performance visibility ``` **Which to choose**: ``` Choose EOA if: - First deployment (testing live behavior) - Running single strategy - Want simple setup - Don't need multiple simultaneous strategies Choose Vault if: - Running multiple strategies - Want professional setup - Need public performance tracking - Trading with significant capital - Building track record ``` ### Leverage Guidelines **Understanding leverage**: ``` Leverage = Position size / Available capital 1x leverage: $1000 capital → $1000 position 2x leverage: $1000 capital → $2000 position 3x leverage: $1000 capital → $3000 position Key points: - Leverage multiplies BOTH gains AND losses - Higher leverage = higher risk - Liquidation risk increases with leverage - Start conservative (1-2x) ``` **Recommended leverage by risk profile**: ``` Conservative (1x): - No amplification - Lower returns, lower risk - Recommended for first deployments - Drawdown ≈ backtest drawdown Moderate (2-3x): - 2-3× returns and risk - Requires careful monitoring - Only after 1x deployment validated - Drawdown ≈ 2-3× backtest drawdown Aggressive (4-5x): - 4-5× returns and risk - Very risky, high liquidation chance - NOT recommended for most users - Drawdown ≈ 4-5× backtest drawdown - Can lose entire capital quickly ``` **Leverage and drawdown**: ``` Backtest: 15% max drawdown 1x deployment: 15% expected drawdown 2x deployment: 30% expected drawdown (may hit margin call) 3x deployment: 45% expected drawdown (very likely liquidation) Rule: Keep leverage low enough that backtest drawdown × leverage < 25% ``` ### Risk Management in Live Trading **Position sizing**: ``` Strategy controls position size via code (85-95% margin usage) Deployment leverage multiplies available margin Total risk = Strategy position size × Deployment leverage Example: - Capital: $1000 - Strategy uses 90% margin - Deployment leverage: 2x - Actual position: $1000 × 0.90 × 2 = $1800 Position size is LARGER than capital (risk of liquidation) ``` **Mental stop loss** (define BEFORE deploying): ``` Example thresholds: - Stop if down 10% from starting capital - Stop if down 15% from peak - Stop if drawdown >1.5× backtest max drawdown Write down your threshold: "I will stop this deployment if capital drops to $______" DO NOT move this threshold once deployed (discipline is critical) ``` **Monitoring frequency**: ``` First 24-48 hours: Check every 2-4 hours First week: Check daily minimum First month: Check every 2-3 days After 1 month: Weekly check acceptable (if performing well) NEVER: - Deploy and forget - Ignore for >1 week during first month - Assume backtest = live performance ``` ## Pre-Deployment Checklist **Complete ALL items before deploying**: ### Strategy Validation - [ ] **Backtested on 6+ months** (12+ months preferred) - [ ] **Sharpe ratio >1.0** (preferably >1.5) - [ ] **Max drawdown <20%** (acceptable risk level) - [ ] **Win rate 45-65%** (realistic range) - [ ] **Profit factor >1.5** (sufficient edge) - [ ] **50+ trades in test** (statistical significance) - [ ] **Multi-period validation** (consistent across different time ranges) - [ ] **Out-of-sample test passed** (performed well on unseen data) ### Code and Logic Review - [ ] **Strategy code reviewed** (no obvious bugs) - [ ] **No look-ahead bias** (not using future data) - [ ] **Indicators validated** (all indicators available and correct) - [ ] **Risk management present** (stop loss and position sizing) - [ ] **Realistic assumptions** (fees, slippage accounted for) ### Operational Readiness - [ ] **Credit balance sufficient** (check with `get_credit_balance`) - [ ] **Deployment type selected** (EOA vs Vault) - [ ] **Leverage set conservatively** (1-2x for first deployment) - [ ] **Monitoring plan established** (how often will you check?) - [ ] **Exit criteria defined** (when will you stop?) - [ ] **Starting capital decided** (how much to deploy?) - [ ] **Capital is risk capital** (can afford to lose 100%) **IF ANY ITEM UNCHECKED: DO NOT DEPLOY** ## Deployment Best Practices ### Start Small **Initial deployment sizing**: ``` WRONG approach: - Backtest shows 50% annual return - Deploy $10,000 immediately - If strategy fails, lose significant capital RIGHT approach: - Deploy $500-1000 initially (5-10% of intended size) - Monitor for 1-2 weeks - Validate live behavior matches backtest - If successful, scale up gradually - Reduce risk during validation phase Scaling schedule example: Week 1-2: $1,000 (test) Week 3-4: $2,000 (if performing well) Week 5-6: $4,000 (if still performing well) Month 2+: Scale to full size gradually ``` **Why start small**: - Live market is different from backtest - Slippage may be higher - Execution may differ - Bugs may only appear in live trading - Can stop with minimal loss if issues arise ### Monitoring Protocol **What to track**: ``` 1. P&L vs backtest expectation: - Is live performance similar to backtest? - Track daily, weekly, monthly returns - Compare to backtest metrics 2. Drawdown: - Current drawdown from peak - Compare to backtest max drawdown - If exceeds backtest max × 1.5, be concerned 3. Trade execution: - Are trades executing as expected? - Check fill prices (slippage) - Verify trade frequency matches backtest 4. Win rate and profit factor: - Track live win rate - Should be close to backtest win rate - If diverges >20%, investigate 5. Market regime: - Has market character changed? - Trending → ranging or vice versa - Strategy may stop working if regime shifts ``` **Daily monitoring checklist (first week)**: - [ ] Check P&L (profit/loss today) - [ ] Check position status (in trade or flat?) - [ ] Check recent trades (executed as expected?) - [ ] Check drawdown (within acceptable range?) - [ ] Note any unusual behavior ### Red Flags - Stop Deployment Immediately **STOP deployment if ANY of these occur**: **1. Excessive drawdown**: ``` Live drawdown >30% OR >1.5× backtest max drawdown Example: - Backtest max drawdown: 15% - Threshold to stop: 22.5% (1.5× backtest) - Current live drawdown: 25% → STOP IMMEDIATELY Why: Strategy may be broken or market changed ``` **2. Win rate collapse**: ``` Live win rate <50% of backtest win rate Example: - Backtest win rate: 55% - Threshold to stop: 27.5% (50% of backtest) - Live win rate after 20 trades: 25% → STOP IMMEDIATELY Why: Strategy logic not working in live market ``` **3. Unexpected trade frequency**: ``` Much higher or lower trade frequency than backtest Example: - Backtest: 2-3 trades per day - Live: 15 trades per day → STOP IMMEDIATELY Why: Strategy may be malfunctioning ``` **4. Consistent losses**: ``` 10+ consecutive losing trades (when backtest shows max 5-6) → STOP IMMEDIATELY Why: Strategy edge may have disappeared ``` **5. Technical issues**: ``` - Orders not executing - Repeated API errors - Position sizing errors - Strategy crashes/restarts frequently → STOP IMMEDIATELY Why: Technical problems = unpredictable risk ``` **6. Market regime change**: ``` Market conditions fundamentally different from backtest period Examples: - Extreme volatility event (>3× normal) - Major regulatory news - Exchange issues → STOP, REASSESS, decide if/when to restart Why: Strategy designed for different conditions ``` ### Post-Deployment Analysis **After 1 week of live trading**: ``` 1. Compare metrics: | Metric | Backtest | Live | Variance | |----------------|----------|-------|----------| | Sharpe | 1.5 | 1.3 | -13% | | Drawdown | 12% | 15% | +25% | | Win rate | 52% | 49% | -6% | | Profit factor | 1.8 | 1.6 | -11% | 2. Evaluate variance: - Small variance (<20%) → Expected, continue ✓ - Moderate variance (20-40%) → Monitor closely, may be temporary - Large variance (>40%) → Significant concern, consider stopping 3. Decision: - If metrics acceptable: Continue monitoring - If metrics concerning: Investigate cause - If red flags present: Stop deployment ``` **After 1 month**: ``` Review: - Total return vs expectation - Max drawdown experienced - Trade execution quality - Any technical issues Decide: - Scale up capital (if performing well) - Continue same size (if acceptable) - Scale down or stop (if underperforming) ``` ## Common Workflows ### Workflow 1: First Deployment (EOA) **Goal**: Deploy strategy for first time to validate live behavior ``` 1. Final pre-deployment check: ☑ Backtested 6+ months (Sharpe 1.4, drawdown 14%) ☑ Code reviewed (no bugs found) ☑ Risk management validated ☑ Starting capital: $500 (can afford to lose) ☑ Monitoring plan: Check daily for first week ☑ Exit criteria: Stop if down >20% or drawdown >25% 2. Check credit balance: get_credit_balance() → Balance: 100 USDC ✓ (sufficient for deployment $0.50) 3. Deploy: deployment_create( strategy_name="RSIMeanReversion_M", symbol="BTC-USDT", timeframe="1h", leverage=1, # Conservative for first deployment deployment_type="eoa" ) → Deployment ID: abc123 → Status: Active → Cost: $0.50 4. Monitor closely: Day 1: Check every 4 hours Day 2-7: Check daily Track: P&L, drawdown, trade execution 5. After 1 week: Review performance vs backtest If good: Continue and consider scaling up If poor: Stop and analyze what went wrong ``` **Cost**: $0.50 ### Workflow 2: Managing Multiple Strategies (Vault) **Goal**: Deploy multiple strategies using Hyperliquid Vault ``` 1. Setup vault (one-time): - Verify 200+ USDC in wallet - Decide vault name (unique, descriptive) 2. Deploy first strategy: deployment_create( strategy_name="TrendFollower_M", symbol="BTC-USDT", timeframe="4h", leverage=2, deployment_type="vault", vault_name="AlgoTrading_Vault_2025", vault_description="Multi-strategy algorithmic trading vault" ) → Vault created successfully 3. Deploy second strategy (same vault): deployment_create( strategy_name="MeanReversion_L", symbol="ETH-USDT", timeframe="1h", leverage=1, deployment_type="vault", vault_name="AlgoTrading_Vault_2025" # Same vault name ) 4. Monitor all deployments: deployment_list() → Shows both strategies with individual performance 5. Manage independently: - Can stop one strategy without affecting other - Each strategy tracks separate P&L - Vault shows combined performance ``` **Cost**: $0.50 per deployment = $1.00 total ### Workflow 3: Stopping Underperforming Deployment **Goal**: Stop deployment when red flags appear ``` 1. Monitor deployment: deployment_list() → Strategy: MomentumBreakout_H → P&L: -18% (started $1000, now $820) → Drawdown: 28% → Red flag: Drawdown > 1.5× backtest max (15% × 1.5 = 22.5%) 2. Decision: STOP (red flag triggered) 3. Stop deployment: deployment_stop(deployment_id="abc123") → Status: Stopped → Final P&L: -$180 (-18%) 4. Analyze what went wrong: - Review trade history - Check market conditions during deployment - Compare to backtest assumptions - Identify issue (market regime change? bug? bad luck?) 5. Next steps: - Fix issues if identified (use improve-trading-strategies) - Re-backtest with improvements - Deploy again with smaller capital if confident - Or abandon strategy if fundamentally broken ``` **Cost**: Free to stop ### Workflow 4: Restarting After Market Change **Goal**: Restart deployment after temporary stop ``` 1. Previously stopped deployment due to high volatility event Stopped during extreme market conditions 2. Market stabilizes: - Check current market conditions - Compare to backtest environment - Decide conditions are favorable again 3. Review strategy: - Re-backtest on recent data - Verify strategy still works - Check no code changes needed 4. Restart deployment: deployment_start(deployment_id="abc123") → Status: Active (resumed) 5. Monitor closely: - First day: Check multiple times - Verify execution matches expectations - Be ready to stop again if issues recur ``` **Cost**: Free ## Troubleshooting ### "Insufficient Credits" **Issue**: Cannot create deployment (balance too low) **Solutions**: ``` 1. Check balance: get_credit_balance() → Balance: 0.20 USDC 2. Purchase credits: - Visit Robonet dashboard - Add credits to account - Deployment costs $0.50 3. Retry deployment after purchase ``` ### "Max 1 EOA Deployment" **Issue**: Trying to create second EOA deployment **Solutions**: ``` 1. Stop existing EOA deployment: deployment_list() → Find existing deployment deployment_stop(deployment_id="existing_id") 2. Or switch to Hyperliquid Vault: - Requires 200+ USDC in wallet - Allows unlimited deployments - Use deployment_type="vault" 3. Or use different wallet (new EOA) ``` ### "Vault Creation Failed" **Issue**: Cannot create Hyperliquid Vault **Solutions**: ``` 1. Verify 200+ USDC in wallet: - Check wallet balance on Hyperliquid - Vault requires minimum balance 2. Check vault name unique: - Try different vault name - Vault names must be unique across Hyperliquid 3. Verify wallet permissions: - Ensure wallet connected properly - Check Hyperliquid account status ``` ### "Live Performance Much Worse Than Backtest" **Issue**: Strategy profitable in backtest, losing in live **Common causes & solutions**: ``` 1. Slippage higher than expected: - Market less liquid than backtest assumed - Solution: Use wider stops, lower frequency trades, or stop deployment 2. Fees not properly accounted: - Forgot to include fees in backtest - Solution: Re-backtest with realistic fees (0.05-0.1%) 3. Market regime changed: - Trending market → ranging market - Solution: Strategy may not work in current conditions, stop deployment 4. Execution delays: - Live trades execute slower than backtest assumed - Solution: Use longer timeframes (1h instead of 5m) 5. Overfitted strategy: - Strategy memorized past data - Solution: Simplify strategy, re-backtest, test on out-of-sample data Decision: If performance -30% worse than backtest, STOP and fix issues ``` ## Legal & Compliance **Important disclaimers**: ``` ⚠️ Trading crypto perpetuals is HIGH RISK ⚠️ Regulations vary by jurisdiction ⚠️ You are responsible for compliance with local laws ⚠️ This is NOT financial advice ⚠️ Trade at your own risk ⚠️ Only risk capital you can afford to lose 100% ``` **User responsibilities**: - Verify trading is legal in your jurisdiction - Understand tax implications of trading - Report trading activity as required by law - Comply with local financial regulations - Maintain records of trading activity **Platform disclaimers**: - Robonet provides tools, not financial advice - Past performance doesn't guarantee future results - No warranty on strategy performance - User bears all risk of capital loss ## Next Steps **If deployment is performing well**: - Continue monitoring regularly - Track performance vs backtest expectations - Consider gradual capital scaling after 1 month - Document what's working for future strategies **If deployment is underperforming**: - Use `improve-trading-strategies` skill to refine - Re-backtest improvements thoroughly - Test with small capital again before scaling **After successful deployment**: - Share learnings (what worked, what didn't) - Consider deploying additional strategies - Build track record for future deployments ## Summary This skill provides **live trading deployment and management**: - **6 tools**: deployment_create ($0.50), deployment_list/start/stop (free), account tools (free) - **Risk**: HIGH (real capital at risk) - **Purpose**: Deploy validated strategies to live trading **Core principle**: Thorough preparation → small initial deployment → active monitoring → gradual scaling. Never deploy without extensive backtesting and clear exit criteria. **Critical warnings**: - You can lose ALL deployed capital - Backtest ≠ live performance (expect differences) - Start small ($500-1000) to validate live behavior - Monitor daily for first week, weekly thereafter - Stop immediately if red flags appear (drawdown >1.5× backtest, win rate collapses, technical issues) - Define exit criteria BEFORE deploying (don't move goalposts) **Pre-deployment checklist must be 100% complete**: Backtest >6 months, Sharpe >1.0, drawdown <20%, code reviewed, monitoring plan, exit criteria, starting small, risk capital only. **Best practice**: Treat first deployment as validation phase, not profit phase. Goal is to confirm strategy works live, not to make money immediately. Profits come after validation succeeds. **Remember**: This is real money, real risk, real consequences. If uncomfortable with any aspect of deployment, DON'T DEPLOY. It's better to miss opportunity than lose capital.