> ## Documentation Index > Fetch the complete documentation index at: https://mathematicalcompany.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # OpenClaw Skill > Trade prediction markets from WhatsApp, Telegram, Discord, and Slack through the OpenClaw bot framework. [OpenClaw](https://github.com/openclaw) is a multi-platform Claude bot framework that runs on WhatsApp, Telegram, Discord, and Slack. The Horizon skill lets users trade prediction markets through natural conversation on any of these platforms. ## How It Works The OpenClaw skill provides a CLI script that Claude calls to interact with the Horizon engine: ``` User (WhatsApp/Telegram/Discord) │ ▼ OpenClaw Bot (Claude) │ reads SKILL.md instructions ▼ scripts/horizon.py [args] │ ▼ horizon.tools → Engine (Rust) ``` Each command prints JSON to stdout. Claude interprets the JSON and responds in natural language. ## Installation ### Option A: Install from ClawHub (recommended) The fastest way to add the Horizon trader skill to your OpenClaw bot: ```bash theme={null} clawhub install horizon-trader ``` Or with npx (no global install needed): ```bash theme={null} npx clawhub@latest install horizon-trader ``` ```bash theme={null} pip install horizon-sdk # or uv pip install horizon-sdk ``` In your OpenClaw `.env` or environment: ```bash theme={null} HORIZON_API_KEY=your-key HORIZON_EXCHANGE=paper ``` The bot will discover the `horizon-trader` skill from `SKILL.md` on startup. ### Option B: Manual install from source ```bash theme={null} pip install horizon-sdk # or uv pip install horizon-sdk ``` ```bash theme={null} cp -r skills/horizon-trader/ /path/to/openclaw/skills/horizon-trader/ ``` Or if you've installed Horizon as a package, the skill is at: ```bash theme={null} cp -r $(python -c "import horizon; print(horizon.__path__[0])")/../skills/horizon-trader /path/to/openclaw/skills/ ``` In your OpenClaw `.env` or environment: ```bash theme={null} HORIZON_API_KEY=your-key HORIZON_EXCHANGE=paper ``` The bot will discover the `horizon-trader` skill from `SKILL.md` on startup. ## Commands The skill exposes these commands via the CLI script: ### Portfolio ```bash theme={null} # Engine status python3 scripts/horizon.py status # List positions python3 scripts/horizon.py positions # List open orders python3 scripts/horizon.py orders # List orders for a specific market python3 scripts/horizon.py orders will-btc-hit-100k # Recent fills python3 scripts/horizon.py fills ``` ### Trading ```bash theme={null} # Submit a limit buy python3 scripts/horizon.py quote will-btc-hit-100k buy 0.55 10 # Submit a limit sell python3 scripts/horizon.py quote will-btc-hit-100k sell 0.60 5 # Cancel an order python3 scripts/horizon.py cancel p1 # Cancel all orders python3 scripts/horizon.py cancel-all # Cancel all orders for a specific market python3 scripts/horizon.py cancel-market will-btc-hit-100k ``` ### Discovery ```bash theme={null} # Search Polymarket python3 scripts/horizon.py discover polymarket "bitcoin" # Search Kalshi python3 scripts/horizon.py discover kalshi "KXBTC" 5 ``` ### Kelly Sizing ```bash theme={null} # kelly [fraction] [max_size] python3 scripts/horizon.py kelly 0.65 0.50 1000 python3 scripts/horizon.py kelly 0.70 0.55 2000 0.5 50 ``` ### Risk Management ```bash theme={null} # Emergency stop python3 scripts/horizon.py kill-switch on "market crash" # Resume trading python3 scripts/horizon.py kill-switch off # Add stop-loss: stop-loss python3 scripts/horizon.py stop-loss will-btc-hit-100k yes sell 10 0.40 # Add take-profit: take-profit python3 scripts/horizon.py take-profit will-btc-hit-100k yes sell 10 0.80 ``` ### Feed Data & Health ```bash theme={null} # Get a feed snapshot python3 scripts/horizon.py feed btc # List all feeds python3 scripts/horizon.py feeds # Check feed staleness and health (optional threshold in seconds) python3 scripts/horizon.py feed-health python3 scripts/horizon.py feed-health 60 ``` ### Contingent Orders ```bash theme={null} # List pending SL/TP python3 scripts/horizon.py contingent # Check parity python3 scripts/horizon.py parity will-btc-hit-100k # Feed connection metrics (single feed or all) python3 scripts/horizon.py feed-metrics btc python3 scripts/horizon.py feed-metrics ``` ### Event Discovery ```bash theme={null} # Discover multi-outcome events on Polymarket python3 scripts/horizon.py discover-events "election" # Top markets by volume python3 scripts/horizon.py top-markets polymarket 10 python3 scripts/horizon.py top-markets kalshi 5 "KXBTC" ``` ### Wallet Analytics (Polymarket - no auth required) ```bash theme={null} # Trade history for a wallet python3 scripts/horizon.py wallet-trades 0x1234... # Open positions with PnL python3 scripts/horizon.py wallet-positions 0x1234... 50 CURRENT # Portfolio value python3 scripts/horizon.py wallet-value 0x1234... # Public profile python3 scripts/horizon.py wallet-profile 0x1234... # Top holders in a market python3 scripts/horizon.py top-holders 0xabc... # Trade flow analysis python3 scripts/horizon.py market-flow 0xabc... 500 ``` ### Hedging & Cross-Market ```bash theme={null} # Compute hedge ratio, effectiveness, basis risk python3 scripts/horizon.py hedge-analysis '[100,102,101,103]' '[50,51,50.5,51.5]' 60 # Scenario analysis on a hedged position python3 scripts/horizon.py hedge-scenario 100 50.0 -50 25.0 '[-0.20,-0.10,0.0,0.10,0.20]' # Multi-ticker hedge: optimal weights python3 scripts/horizon.py hedge-multi '[0.01,-0.02,0.03]' '[[0.005,-0.01,0.015],[0.008,-0.015,0.02]]' '["SPY","QQQ"]' # Hedge cost summary python3 scripts/horizon.py hedge-costs ``` ### Account & Credentials ```bash theme={null} # Always pass password via env var (avoids shell history exposure) # Create account and generate SDK key (encrypted, saved to ~/.horizon/credentials.json) HORIZON_PASSWORD="mypassword" python3 scripts/horizon.py setup user@example.com "My Name" # Login and generate new SDK key HORIZON_PASSWORD="mypassword" python3 scripts/horizon.py login user@example.com # Check credential status python3 scripts/horizon.py key-status ``` ### Simulation & Arbitrage ```bash theme={null} # Monte Carlo on current positions python3 scripts/horizon.py simulate 50000 # Cross-exchange arb python3 scripts/horizon.py arb will-btc-hit-100k kalshi polymarket 0.48 0.52 10 ``` ## Conversation Examples ### Telegram ``` User: What's my PnL today? Bot: 📊 Your engine status: Daily PnL: +$9.25 Realized: +$12.45 | Unrealized: -$3.20 3 open orders, 2 positions User: Find me some bitcoin markets Bot: Found 5 markets on Polymarket: 1. Will BTC hit $100k by end of 2025? 2. Will BTC hit $150k by end of 2025? ... User: Buy 20 contracts of will-btc-hit-100k at 0.52 Bot: ✅ Order submitted: Order ID: p1 Buy 20 contracts @ 0.52 Market: will-btc-hit-100k ``` ### WhatsApp ``` User: kill switch NOW Bot: 🚨 Kill switch activated. All orders canceled. Reason: user emergency request Send "resume trading" to deactivate. ``` ## SKILL.md Reference The `SKILL.md` file tells Claude when and how to use the skill: ```yaml theme={null} --- name: horizon-trader description: Trade prediction markets (Polymarket, Kalshi) emoji: "\U0001F4C8" metadata: requires_env: - HORIZON_API_KEY install: pip install horizon-sdk --- ``` The body contains natural language instructions for: * When to activate (user asks about trading, positions, markets) * Command syntax with examples * Output format (JSON) * Safety notes (confirm before submitting orders) ## Output Format All commands return JSON. On success: ```json theme={null} { "running": true, "kill_switch_active": false, "open_orders": 3, "daily_pnl": 9.25 } ``` On error: ```json theme={null} { "error": "risk rejection: max position exceeded" } ``` The bot interprets the JSON and responds in natural language, never showing raw JSON to the user. ## Pipeline Features (v0.4.0) The Horizon SDK includes advanced pipeline components for automated strategies: * **Regime Detection** (`regime_signal`) - volatility/trend regime classification * **Feed Guard** (`feed_guard`) - auto-activates kill switch when feeds go stale * **Inventory Skew** (`inventory_skewer`) - shifts quotes to reduce position risk * **Adaptive Spread** (`adaptive_spread`) - dynamically adjusts spread based on market conditions * **Execution Tracker** (`execution_tracker`) - monitors fill quality and adverse selection * **Multi-Strategy** - run different pipelines per market via dict config * **Cross-Market Hedging** (`cross_hedger`) - generates hedge quotes when portfolio delta exceeds threshold These are used with `hz.run()` in Python strategies. See the [full documentation](https://docs.openclaw.ai/tools/clawhub) for details. ## Deployment Deploy the OpenClaw skill with the Horizon trader to your preferred platform. Full deployment guide: [https://docs.openclaw.ai/tools/clawhub](https://docs.openclaw.ai/tools/clawhub)