Open-broker

Hyperliquid trading plugin with background position monitoring and custom automations. Execute market orders, limit orders, manage positions, view funding ra...

Security Scan

The skill's requested binaries, npm install, and HYPERLIQUID_PRIVATE_KEY credential align with a Hyperliquid trading CLI — but it will handle private keys and persist wallet config on the host, so treat credentials and installs carefully.

Purpose & Capability

Name/description, required binary 'openbroker', npm install package 'openbroker', and the HYPERLIQUID_PRIVATE_KEY credential are coherent for a trading CLI that places orders and monitors positions.

Instruction Scope

SKILL.md instructs the agent to run the openbroker CLI (setup, buy/sell, search, etc.), and explicitly advises falling back to Bash/CLI for missing plugin-tool functionality and long-running strategies. This is within the stated purpose but gives the agent broad ability to run CLI commands that create wallets, save config files, and access private keys.

Install Mechanism

Installation is via the npm package 'openbroker' (creates the openbroker binary). This is a standard registry install (moderate trust cost but expected for a Node CLI). No ad-hoc downloads or unknown URLs are used.

Credentials

Only HYPERLIQUID_PRIVATE_KEY is required and declared as primaryEnv, which is appropriate for trading operations. Note: the CLI also supports generating and storing wallets/config locally and an 'API wallet' flow (which the skill documents), so the skill may create or persist additional secret material on disk.

Persistence & Privilege

The skill is not force-enabled (always:false) and supports autonomous invocation (normal). However, the documented 'setup' flow writes wallet/config files and auto-approves builder fees — the skill therefore persists credentials/config to the host filesystem (expected for a trading CLI but a material privilege).

Assessment

This skill appears to be what it says: a CLI-based Hyperliquid trading tool that needs a Hyperliquid private key. Before installing, consider: (1) Only provide a restricted API wallet/private key (one that cannot withdraw) if possible; avoid giving a full master private key. (2) Inspect the npm package (source repo / package contents) or install in an isolated/sandbox environment. (3) Be aware 'openbroker setup' can generate and save wallet keys/config to disk — locate and protect or remove those files after use. (4) Prefer the documented API-wallet (no-withdraw) mode or use --dry for previews. (5) If you do not trust running shell commands from the agent, do not enable CLI-fallback behavior or run the skill in a constrained environment.

Open Broker - Hyperliquid Trading CLI

Execute trading operations on Hyperliquid DEX with builder fee support.

Installation

npm install -g openbroker

Quick Start

# 1. Setup (generates wallet, creates config, approves builder fee)
openbroker setup

# 2. Fund your wallet with USDC on Arbitrum, then deposit at https://app.hyperliquid.xyz/

# 3. Start trading
openbroker account
openbroker buy --coin ETH --size 0.1

Important: Finding Assets Before Trading

Always search before trading an unfamiliar asset. Hyperliquid has main perps (ETH, BTC, SOL...), HIP-3 perps (xyz:CL, xyz:GOLD, km:USOIL...), and spot markets. Use search to discover the correct ticker:

openbroker search --query GOLD              # Find all GOLD markets across all providers
openbroker search --query oil               # Find oil-related assets (CL, BRENTOIL, USOIL...)
openbroker search --query BTC --type perp   # BTC perps only
openbroker search --query NATGAS --type hip3  # HIP-3 only

Or with the ob_search plugin tool: { "query": "gold" } or { "query": "oil", "type": "hip3" }

HIP-3 assets use dex:COIN format — e.g., xyz:CL not just CL. If you get an error like "No market data found", search for the asset to find the correct prefixed ticker. Common HIP-3 dexes: xyz, flx, km, hyna, vntl, cash.

Troubleshooting: CLI Fallback

If an ob_* plugin tool returns unexpected errors, empty results, or crashes, fall back to the equivalent CLI command via Bash. The CLI and plugin tools share the same core code, but the CLI has more mature error handling and output.

Plugin Tool	CLI Equivalent
ob_account	openbroker account --json
ob_positions	openbroker positions --json
ob_funding	openbroker funding --json --include-hip3
ob_markets	openbroker markets --json --include-hip3
ob_search	openbroker search --query <QUERY>
ob_buy	openbroker buy --coin <COIN> --size <SIZE>
ob_sell	openbroker sell --coin <COIN> --size <SIZE>
ob_limit	openbroker limit --coin <COIN> --side <SIDE> --size <SIZE> --price <PRICE>
ob_tpsl	openbroker tpsl --coin <COIN> --tp <PRICE> --sl <PRICE>
ob_cancel	openbroker cancel --all or --coin <COIN>
ob_fills	openbroker fills --json
ob_orders	openbroker orders --json
ob_funding_scan	openbroker funding-scan --json
ob_candles	openbroker candles --coin <COIN> --json
ob_auto_run	openbroker auto run <script> [--dry]
ob_auto_stop	(stop via SIGINT when using CLI)
ob_auto_list	openbroker auto list

When to use CLI fallback:

• Plugin tool returns null, empty data, or throws an error
• You need data the plugin tool doesn't expose (e.g., --verbose debug output)
• Long-running operations (strategies) — the CLI handles timeouts and progress better

Add --dry to any trading CLI command to preview without executing. Add --json to info commands for structured output.

Command Reference

Setup

openbroker setup              # One-command setup (wallet + config + builder approval)
openbroker approve-builder --check  # Check builder fee status (for troubleshooting)

The setup command offers three modes:

1. Generate fresh wallet (recommended for agents) — creates a dedicated trading wallet with builder fee auto-approved. No browser steps needed — just fund with USDC and start trading.
2. Import existing key — use a private key you already have
3. Generate API wallet — creates a restricted wallet that can trade but cannot withdraw. Requires browser approval from a master wallet.

For options 1 and 2, setup saves config and approves the builder fee automatically. For option 3 (API wallet), see the API Wallet Setup section below.

Fresh Wallet Setup (Recommended for Agents)

The simplest setup for agents. A fresh wallet is generated, the builder fee is auto-approved, and the agent is ready to trade immediately after funding.

Flow:

1. Run openbroker setup and choose option 1 ("Generate a fresh wallet")
2. The CLI generates a wallet, saves the config, and approves the builder fee automatically
3. Fund the wallet by sending USDC from your Hyperliquid account to the agent's wallet address using the Send feature on https://app.hyperliquid.xyz/. Funding should be done on Hyperliquid L1 only.
4. Start trading

API Wallet Setup (Alternative)

API wallets can place trades on behalf of a master account but cannot withdraw funds. Use this if you prefer to keep funds in your existing wallet and only delegate trading access.

Flow:

1. Run openbroker setup and choose option 3 ("Generate API wallet")
2. The CLI generates a keypair and prints an approval URL (e.g. https://openbroker.dev/approve?agent=0xABC...)
3. The agent owner opens the URL in a browser and connects their master wallet (MetaMask etc.)
4. The master wallet signs two transactions: ApproveAgent (authorizes the API wallet) and ApproveBuilderFee (approves the 1 bps fee)
5. The CLI detects the approval automatically and saves the config

After setup, the config will contain:

HYPERLIQUID_PRIVATE_KEY=0x...        # API wallet private key
HYPERLIQUID_ACCOUNT_ADDRESS=0x...    # Master account address
HYPERLIQUID_NETWORK=mainnet

Important for agents: When using an API wallet, pass the approval URL to the agent owner (the human who controls the master wallet). The owner must approve in a browser before the agent can trade. The CLI waits up to 10 minutes for the approval. If it times out, re-run openbroker setup.

Account Info

openbroker account            # Balance, equity, margin
openbroker account --orders   # Include open orders
openbroker positions          # Open positions with PnL
openbroker positions --coin ETH  # Specific coin

Funding Rates

openbroker funding --top 20   # Top 20 by funding rate
openbroker funding --coin ETH # Specific coin

Markets

openbroker markets --top 30   # Top 30 main perps
openbroker markets --coin BTC # Specific coin

All Markets (Perps + Spot + HIP-3)

openbroker all-markets                 # Show all markets
openbroker all-markets --type perp     # Main perps only
openbroker all-markets --type hip3     # HIP-3 perps only
openbroker all-markets --type spot     # Spot markets only
openbroker all-markets --top 20        # Top 20 by volume

Search Markets (Find assets across providers)

openbroker search --query GOLD    # Find all GOLD markets
openbroker search --query BTC     # Find BTC across all providers
openbroker search --query ETH --type perp  # ETH perps only

Spot Markets

openbroker spot                   # Show all spot markets
openbroker spot --coin PURR       # Show PURR market info
openbroker spot --balances        # Show your spot balances
openbroker spot --top 20          # Top 20 by volume

Trade Fills

openbroker fills                          # Recent fills
openbroker fills --coin ETH               # ETH fills only
openbroker fills --coin BTC --side buy --top 50

Order History

openbroker orders                         # Recent orders (all statuses)
openbroker orders --open                  # Currently open orders only
openbroker orders --open --coin ETH       # Open orders for a specific coin
openbroker orders --coin ETH --status filled
openbroker orders --top 50

Order Status

openbroker order-status --oid 123456789   # Check specific order
openbroker order-status --oid 0x1234...   # By client order ID

Fee Schedule

openbroker fees                           # Fee tier, rates, and volume

Candle Data (OHLCV)

openbroker candles --coin ETH                           # 24 hourly candles
openbroker candles --coin BTC --interval 4h --bars 48   # 48 four-hour bars
openbroker candles --coin SOL --interval 1d --bars 30   # 30 daily bars

Funding History

openbroker funding-history --coin ETH              # Last 24h
openbroker funding-history --coin BTC --hours 168  # Last 7 days

Recent Trades (Tape)

openbroker trades --coin ETH              # Last 30 trades
openbroker trades --coin BTC --top 50     # Last 50 trades

Rate Limit

openbroker rate-limit                     # API usage and capacity

Funding Rate Scanner (Cross-Dex)

openbroker funding-scan                          # Scan all dexes, >25% threshold
openbroker funding-scan --threshold 50 --pairs   # Show opposing funding pairs
openbroker funding-scan --hip3-only --top 20     # HIP-3 only
openbroker funding-scan --watch --interval 120   # Re-scan every 2 minutes

Trading Commands

HIP-3 Perp Trading

All trading commands support HIP-3 assets using dex:COIN syntax:

openbroker buy --coin xyz:CL --size 1              # Buy crude oil on xyz dex
openbroker sell --coin xyz:BRENTOIL --size 1        # Sell brent oil
openbroker limit --coin xyz:GOLD --side buy --size 0.1 --price 2500

Market Orders (Quick)

openbroker buy --coin ETH --size 0.1
openbroker sell --coin BTC --size 0.01
openbroker buy --coin SOL --size 5 --slippage 100  # Custom slippage (bps)

Market Orders (Full)

openbroker market --coin ETH --side buy --size 0.1
openbroker market --coin BTC --side sell --size 0.01 --slippage 100

Limit Orders

openbroker limit --coin ETH --side buy --size 1 --price 3000
openbroker limit --coin SOL --side sell --size 10 --price 200 --tif ALO

Set TP/SL on Existing Position

# Set take profit at $40, stop loss at $30
openbroker tpsl --coin HYPE --tp 40 --sl 30

# Set TP at +10% from entry, SL at entry (breakeven)
openbroker tpsl --coin HYPE --tp +10% --sl entry

# Set only stop loss at -5% from entry
openbroker tpsl --coin ETH --sl -5%

# Partial position TP/SL
openbroker tpsl --coin ETH --tp 4000 --sl 3500 --size 0.5

Trigger Orders (Standalone TP/SL)

# Take profit: sell when price rises to $40
openbroker trigger --coin HYPE --side sell --size 0.5 --trigger 40 --type tp

# Stop loss: sell when price drops to $30
openbroker trigger --coin HYPE --side sell --size 0.5 --trigger 30 --type sl

Cancel Orders

openbroker cancel --all           # Cancel all orders
openbroker cancel --coin ETH      # Cancel ETH orders only
openbroker cancel --oid 123456    # Cancel specific order

Advanced Execution

TWAP (Native Exchange Order)

# Buy 1 ETH over 30 minutes (exchange handles slicing)
openbroker twap --coin ETH --side buy --size 1 --duration 30

# Sell 0.5 BTC over 2 hours without randomized timing
openbroker twap --coin BTC --side sell --size 0.5 --duration 120 --randomize false

# Cancel a running TWAP
openbroker twap-cancel --coin ETH --twap-id 77738308

# Check TWAP status
openbroker twap-status --active

Scale In/Out (Grid Orders)

# Place 5 buy orders ranging 2% below current price
openbroker scale --coin ETH --side buy --size 1 --levels 5 --range 2

# Scale out with exponential distribution
openbroker scale --coin BTC --side sell --size 0.5 --levels 4 --range 3 --distribution exponential --reduce

Bracket Order (Entry + TP + SL)

# Long ETH with 3% take profit and 1.5% stop loss
openbroker bracket --coin ETH --side buy --size 0.5 --tp 3 --sl 1.5

# Short with limit entry
openbroker bracket --coin BTC --side sell --size 0.1 --entry limit --price 100000 --tp 5 --sl 2

Chase Order (Follow Price)

# Chase buy with ALO orders until filled
openbroker chase --coin ETH --side buy --size 0.5 --timeout 300

# Aggressive chase with tight offset
openbroker chase --coin SOL --side buy --size 10 --offset 2 --timeout 60

Order Types

Limit Orders vs Trigger Orders

Limit Orders (openbroker limit):

• Execute immediately if price is met
• Rest on the order book until filled or cancelled
• A limit sell BELOW current price fills immediately (taker)
• NOT suitable for stop losses

Trigger Orders (openbroker trigger, openbroker tpsl):

• Stay dormant until trigger price is reached
• Only activate when price hits the trigger level
• Proper way to set stop losses and take profits
• Won't fill prematurely

When to Use Each

Scenario	Command
Buy at specific price below market	openbroker limit
Sell at specific price above market	openbroker limit
Stop loss (exit if price drops)	openbroker trigger --type sl
Take profit (exit at target)	openbroker trigger --type tp
Add TP/SL to existing position	openbroker tpsl

Common Arguments

All commands support --dry for dry run (preview without executing).

Argument	Description
--coin	Asset symbol (ETH, BTC, SOL, HYPE, etc.)
--side	Order direction: buy or sell
--size	Order size in base asset
--price	Limit price
--dry	Preview without executing
--help	Show command help

Order Arguments

Argument	Description
--trigger	Trigger price (for trigger orders)
--type	Trigger type: tp or sl
--slippage	Slippage tolerance in bps (for market orders)
--tif	Time in force: GTC, IOC, ALO
--reduce	Reduce-only order

TP/SL Price Formats

Format	Example	Description
Absolute	--tp 40	Price of $40
Percentage up	--tp +10%	10% above entry
Percentage down	--sl -5%	5% below entry
Entry price	--sl entry	Breakeven stop

Configuration

Config is loaded from (in priority order):

1. Environment variables
2. .env in current directory
3. ~/.openbroker/.env (global config)

Run openbroker setup to create the global config interactively.

Variable	Required	Description
HYPERLIQUID_PRIVATE_KEY	Yes	Wallet private key (0x...)
HYPERLIQUID_NETWORK	No	mainnet (default) or testnet
HYPERLIQUID_ACCOUNT_ADDRESS	No	Master account address (required for API wallets)

The builder fee (1 bps / 0.01%) is hardcoded and not configurable.

OpenClaw Plugin (Optional)

This skill works standalone via Bash — every command above runs through the openbroker CLI. For enhanced features, the same openbroker npm package also ships as an OpenClaw plugin that you can enable alongside this skill.

What the plugin adds

• Structured agent tools (ob_account, ob_buy, ob_limit, etc.) — typed tool calls with proper input schemas instead of Bash strings. The agent gets structured JSON responses.
• Background position watcher — polls your Hyperliquid account every 30s and sends webhook alerts when positions open/close, PnL moves significantly, or margin usage gets dangerous.
• Automation tools (ob_auto_run, ob_auto_stop, ob_auto_list) — start, stop, and manage custom trading automations from within the agent.
• CLI commands — openclaw ob status and openclaw ob watch for inspecting the watcher.

Enable the plugin

The plugin is bundled in the same openbroker npm package. To enable it in your OpenClaw config:

plugins:
  entries:
    openbroker:
      enabled: true
      config:
        hooksToken: "your-hooks-secret"   # Required for watcher alerts
        watcher:
          enabled: true
          pollIntervalMs: 30000
          pnlChangeThresholdPct: 5
          marginUsageWarningPct: 80

The plugin reads wallet credentials from ~/.openbroker/.env (set up by openbroker setup), so you don't need to duplicate privateKey in the plugin config unless you want to override.

Webhook setup for watcher alerts

For position alerts to reach the agent, enable hooks in your gateway config:

hooks:
  enabled: true
  token: "your-hooks-secret"   # Must match hooksToken above

Without hooks, the watcher still runs and tracks state (accessible via ob_watcher_status), but it can't wake the agent.

Using with or without the plugin

• Skill only (no plugin): Use Bash commands (openbroker buy --coin ETH --size 0.1). No background monitoring.
• Skill + plugin: The agent prefers the ob_* tools when available (structured data), falls back to Bash for commands not covered by tools (strategies, scale). Background watcher sends alerts automatically.

Trading Automations

Automations let you write custom event-driven trading logic as TypeScript scripts. Write exactly the logic you need and OpenBroker handles the polling, event detection, and SDK access.

IMPORTANT: Bundled Examples Are References, Not Ready-Made Strategies

OpenBroker ships bundled automation examples (dca, grid, funding-arb, mm-spread, mm-maker) that demonstrate common patterns. These examples are meant to be read and learned from — not run directly as production strategies. When the user asks for a strategy (e.g. "run a DCA on HYPE"), you should:

1. Read the relevant bundled example source code to understand the automation API patterns
2. Write a custom automation tailored to the user's specific requirements (coin, size, thresholds, risk parameters, TP/SL logic, etc.)
3. Save it to ~/.openbroker/automations/ and run it with openbroker auto run

Do NOT simply run openbroker auto run --example dca --set coin=HYPE. The examples use generic defaults and lack the nuanced risk management, position sizing, and TP/SL logic that a real strategy needs. Always write a purpose-built automation.

To view bundled examples and their config schemas:

openbroker auto examples    # List examples with config fields

Available examples: dca, grid, funding-arb, mm-spread, mm-maker

How Automations Work

An automation is a .ts file that exports a default function. The function receives an AutomationAPI with the full Hyperliquid client, typed event subscriptions, persistent state, and a logger. The runtime polls Hyperliquid every 10s (configurable) and dispatches events when changes are detected.

Writing an Automation

Create a .ts file in ~/.openbroker/automations/ (or any path):

// ~/.openbroker/automations/funding-scalp.ts
export default function(api) {
  const COIN = 'ETH';

  api.on('funding_update', async ({ coin, annualized }) => {
    if (coin !== COIN) return;

    if (annualized > 0.5 && !api.state.get('isShort')) {
      api.log.info('High positive funding — going short');
      await api.client.marketOrder(COIN, false, 0.1);
      api.state.set('isShort', true);
    } else if (annualized < -0.1 && api.state.get('isShort')) {
      api.log.info('Funding normalized — closing short');
      await api.client.marketOrder(COIN, true, 0.1);
      api.state.set('isShort', false);
    }
  });

  api.onStop(async () => {
    if (api.state.get('isShort')) {
      api.log.warn('Closing short on shutdown');
      await api.client.marketOrder('ETH', true, 0.1);
      api.state.set('isShort', false);
    }
  });
}

AutomationAPI Reference

Property / Method	Description
api.client	Full HyperliquidClient — marketOrder(), limitOrder(), triggerOrder(), cancelAll(), getUserStateAll(), getAllMids(), updateLeverage(), and 35+ more methods
api.on(event, handler)	Subscribe to a market/account event (see Events below)
api.every(ms, handler)	Run a handler on a recurring interval (aligned to poll loop)
api.onStart(handler)	Called after all handlers are registered, before first poll
api.onStop(handler)	Called on shutdown (SIGINT). Use for cleanup — close positions, cancel orders
api.onError(handler)	Called when a handler throws. Error is already logged — use for recovery logic
api.state.get(key)	Get a persisted value (survives restarts, stored in ~/.openbroker/state/)
api.state.set(key, value)	Set a persisted value
api.state.delete(key)	Delete a persisted value
api.state.clear()	Clear all state
api.publish(message, options?)	Send a message to the OpenClaw agent via webhook. Triggers an agent turn — the agent receives the message and can notify the user, take action, etc. Returns true if delivered. Options: { name?, wakeMode?, deliver?, channel? }
api.log.info/warn/error/debug(msg)	Structured logger
api.utils	roundPrice, roundSize, sleep, normalizeCoin, formatUsd, annualizeFundingRate
api.id	Automation ID (filename or --id flag)
api.dryRun	true if running with --dry (write methods are intercepted)

Events

Event	Payload	When
tick	{ timestamp, pollCount }	Every poll cycle (default: 10s)
price_change	{ coin, oldPrice, newPrice, changePct }	Mid price moved > 0.01% between polls
funding_update	{ coin, fundingRate, annualized, premium }	Every poll for all assets
position_opened	{ coin, side, size, entryPrice }	New position detected
position_closed	{ coin, previousSize, entryPrice }	Position no longer present
position_changed	{ coin, oldSize, newSize, entryPrice }	Position size changed
pnl_threshold	{ coin, unrealizedPnl, changePct, positionValue }	PnL moved > 5% of position value
margin_warning	{ marginUsedPct, equity, marginUsed }	Margin usage > 80%

Event Details — Choosing the Right Event

tick — The universal heartbeat

Fires every single poll cycle (default: 10s) regardless of market conditions. Use this when you need to check something on every poll — absolute price thresholds, custom conditions, periodic account checks. This is the most reliable event because it always fires.

Payload: { timestamp: number, pollCount: number }

When to use:

• Checking if a price is above/below an absolute threshold (e.g. "alert me when ETH < $3000")
• Custom conditions that don't fit other events (e.g. "if I have no positions and funding is high, enter")
• Periodic tasks that need to run every poll (though api.every() is better for longer intervals)

Example — absolute price alert:

api.on('tick', async () => {
  const mids = await api.client.getAllMids();
  const price = parseFloat(mids['HYPE']);
  if (price < 38 && !api.state.get('alerted')) {
    api.state.set('alerted', true);
    await api.publish(`HYPE dropped below $38 — now at $${price.toFixed(3)}`);
  }
});

Note: tick does not include price data in its payload — you must fetch it yourself via api.client.getAllMids(). This is because tick fires before any other event processing. If you only care about price movements, use price_change instead.

price_change — Relative price movements

Fires when a coin's mid price moves ≥ 0.01% compared to the previous poll. This filters out rounding noise while catching virtually any real price movement. The comparison is between consecutive polls (not from a fixed baseline), so it detects incremental changes.

Payload: { coin: string, oldPrice: number, newPrice: number, changePct: number }

When to use:

• Reacting to price movements (breakouts, momentum, mean reversion)
• Monitoring specific coins for volatility
• Building price-triggered entry/exit logic

When NOT to use:

• Checking if price is above/below a fixed threshold — use tick instead, because price_change only fires on relative movement between polls. During slow drifts (e.g. price slowly declining $0.001/s), the change between any two 10s polls may be < 0.01%, so the event won't fire even though the price has crossed your threshold.

Example — momentum detector:

api.on('price_change', async ({ coin, changePct, newPrice }) => {
  if (coin !== 'ETH') return;
  if (changePct > 0.5) {
    api.log.info(`ETH surging +${changePct.toFixed(2)}% — price $${newPrice}`);
    // Enter long on strong upward momentum
  }
});

funding_update — Funding rate data

Fires every poll for every asset that has funding rate data. This is high-frequency — if there are 150 perp assets, this fires 150 times per poll. Filter by coin in your handler.

Payload: { coin: string, fundingRate: number, annualized: number, premium: number }

• fundingRate — the raw hourly funding rate (e.g. 0.0001 = 0.01%/hr)
• annualized — annualized rate (fundingRate × 8760 × 100, as a percentage)
• premium — the premium component

When to use:

• Funding rate arbitrage strategies
• Monitoring for extreme funding (entry/exit signals)
• Scanning for highest/lowest funding across all assets

Example — funding scalp:

api.on('funding_update', async ({ coin, annualized }) => {
  if (coin !== 'ETH') return;
  if (annualized > 50 && !api.state.get('isShort')) {
    api.log.info(`ETH funding at ${annualized.toFixed(1)}% annualized — shorting`);
    await api.client.marketOrder('ETH', false, 0.1);
    api.state.set('isShort', true);
  }
});

position_opened — New position detected

Fires when a position appears that wasn't present in the previous poll. Useful for tracking entries made by other systems or confirming your own orders filled.

Payload: { coin: string, side: 'long' | 'short', size: number, entryPrice: number }

When to use:

• Setting TP/SL on new positions automatically
• Logging/alerting when positions are opened (by you or another system)
• Starting position-specific monitoring

Example — auto TP/SL on new positions:

api.on('position_opened', async ({ coin, side, size, entryPrice }) => {
  const tpPrice = side === 'long' ? entryPrice * 1.05 : entryPrice * 0.95;
  const slPrice = side === 'long' ? entryPrice * 0.97 : entryPrice * 1.03;
  await api.client.takeProfit(coin, side !== 'long', size, tpPrice);
  await api.client.stopLoss(coin, side !== 'long', size, slPrice);
  api.log.info(`Set TP at ${tpPrice} / SL at ${slPrice} for ${coin}`);
});

position_closed — Position gone

Fires when a position that existed in the previous poll is no longer present. The position was either closed by you, liquidated, or filled by TP/SL.

Payload: { coin: string, previousSize: number, entryPrice: number }

When to use:

• Logging/alerting when positions close
• Cleaning up related orders or state
• Re-entry logic after a position closes

Example:

api.on('position_closed', async ({ coin, previousSize, entryPrice }) => {
  api.log.info(`${coin} position closed (was ${previousSize} @ ${entryPrice})`);
  api.state.delete(`${coin}_tp`);
  await api.publish(`Position closed: ${coin} (entry: $${entryPrice})`);
});

position_changed — Size or direction changed

Fires when an existing position's size changes (partial close, add to position, or flip direction). Does NOT fire when a new position opens or an existing one fully closes — use position_opened and position_closed for those.

Payload: { coin: string, oldSize: number, newSize: number, entryPrice: number }

• oldSize/newSize are signed: positive = long, negative = short

When to use:

• Detecting partial closes or position scaling
• Adjusting TP/SL when position size changes
• Tracking DCA entries

Example:

api.on('position_changed', async ({ coin, oldSize, newSize }) => {
  if (Math.abs(newSize) > Math.abs(oldSize)) {
    api.log.info(`${coin} position increased: ${oldSize} → ${newSize}`);
  } else {
    api.log.info(`${coin} position reduced: ${oldSize} → ${newSize}`);
  }
});

pnl_threshold — Significant PnL movement

Fires when unrealized PnL changes by ≥ 5% of position value between consecutive polls. This is a large move detector — useful for risk management alerts rather than routine monitoring.

Payload: { coin: string, unrealizedPnl: number, changePct: number, positionValue: number }

• changePct — the PnL change as a percentage of total position value (not % of PnL itself)

When to use:

• Risk alerts for large PnL swings
• Auto-close or reduce positions on sudden adverse moves
• Escalating alerts to the user via api.publish()

Example:

api.on('pnl_threshold', async ({ coin, unrealizedPnl, changePct }) => {
  if (unrealizedPnl < 0) {
    await api.publish(
      `⚠️ ${coin} PnL dropped sharply: $${unrealizedPnl.toFixed(2)} (${changePct.toFixed(1)}% of position)`,
      { name: 'pnl-alert' },
    );
  }
});

margin_warning — High margin usage

Fires when margin usage exceeds 80% of equity. After the first trigger, it only fires again if margin usage increases by another 5 percentage points (prevents spam). Resets when margin drops back below 80%.

Payload: { marginUsedPct: number, equity: number, marginUsed: number }

When to use:

• Automated risk reduction (close smallest position to free margin)
• Alerting the user before liquidation risk
• Pausing new entries when margin is high

Example:

api.on('margin_warning', async ({ marginUsedPct, equity }) => {
  await api.publish(
    `🚨 Margin at ${marginUsedPct.toFixed(1)}% — equity: $${equity.toFixed(2)}. Consider reducing exposure.`,
    { name: 'margin-alert' },
  );
});

Choosing the Right Event — Quick Guide

Use case	Best event	Why
Alert when price crosses a fixed level	tick	Fires every poll — no minimum change threshold
React to price momentum/volatility	price_change	Provides relative change data between polls
Funding rate strategy	funding_update	Gives annualized rate directly
Auto TP/SL on new positions	position_opened	Fires exactly when a new position appears
Log when positions close	position_closed	Fires when position disappears
Track position scaling	position_changed	Fires on size changes only
Risk management — PnL spikes	pnl_threshold	Only fires on large moves (≥5% of position value)
Risk management — margin	margin_warning	Fires at 80%+ margin usage
Periodic task (DCA, rebalance)	api.every(ms, fn)	Better than tick for longer intervals

Client Methods Available

The api.client object exposes the full HyperliquidClient. All coin params accept HIP-3 prefixed tickers (e.g. xyz:CL). Optional user params default to the configured wallet address.

Trading

Method	Description
marketOrder(coin, isBuy, size, slippageBps?, leverage?)	Market order via IOC limit at mid ± slippage. Returns OrderResponse
limitOrder(coin, isBuy, size, price, tif?, reduceOnly?, leverage?)	Limit order. tif: 'Gtc' (default), 'Ioc', 'Alo'. Returns OrderResponse
triggerOrder(coin, isBuy, size, triggerPrice, limitPrice, tpsl, reduceOnly?, leverage?)	Trigger (conditional) order. tpsl: 'tp' or 'sl'. Activates when price hits triggerPrice, then fills as limit at limitPrice. Returns OrderResponse
stopLoss(coin, isBuy, size, triggerPrice, slippageBps?)	Stop loss shortcut. Sets limit price with slippage buffer (default 100 bps / 1%) to ensure fill. reduceOnly is always true. Returns OrderResponse
takeProfit(coin, isBuy, size, triggerPrice)	Take profit shortcut. Limit price = trigger price (favorable direction). reduceOnly is always true. Returns OrderResponse
cancel(coin, oid)	Cancel a single order by numeric OID. Returns CancelResponse
cancelAll(coin?)	Cancel all open orders. If coin is provided, only cancels orders for that asset. Returns CancelResponse[]
order(coin, isBuy, size, price, orderType, reduceOnly?, includeBuilder?, leverage?)	Low-level order placement. orderType: `{ limit: { tif: 'Gtc'

Market Data

Method	Returns
getAllMids()	Record<string, string> — mid prices for all assets (main + HIP-3). Key = coin name, value = price string
getMetaAndAssetCtxs()	MetaAndAssetCtxs — market metadata (universe of assets with szDecimals, maxLeverage) and asset contexts (funding, open interest, volume, mark/oracle prices)
getL2Book(coin)	{ bids, asks, bestBid, bestAsk, midPrice, spread, spreadBps } — L2 order book with computed spread
getRecentTrades(coin)	Array<{ coin, side, px, sz, time, hash, tid }> — recent trade tape. side: 'B' (buy) or 'A' (sell)
getCandleSnapshot(coin, interval, startTime, endTime?)	Array<{ t, T, s, i, o, c, h, l, v, n }> — OHLCV candles. interval: '1m', '5m', '15m', '1h', '4h', '1d'. Times are Unix ms
getFundingHistory(coin, startTime, endTime?)	Array<{ coin, fundingRate, premium, time }> — historical hourly funding rates
getPredictedFundings()	Array<[coin, Array<[venue, { fundingRate, nextFundingTime }]>]> — predicted funding rates across all venues
getPerpDexs()	`Array<{ name, fullName, deployer }
getAllPerpMetas()	Array<{ dexName, meta, assetCtxs }> — metadata + contexts for every perp DEX (main + all HIP-3)
getSpotMeta()	{ tokens, universe } — spot market metadata (token info, trading pairs)
getSpotMetaAndAssetCtxs()	{ meta, assetCtxs } — spot metadata + price/volume contexts
getTokenDetails(tokenId)	Token details: supply, deployer, prices. Returns null if not found

Account

Method	Returns
getUserStateAll(user?)	ClearinghouseState — full account state across all dexes: marginSummary (accountValue, totalMarginUsed, withdrawable), crossMarginSummary, and assetPositions[] (each with position.coin, .szi, .entryPx, .unrealizedPnl, .positionValue, .leverage, .marginUsed, .liquidationPx)
getUserState(user?, dex?)	ClearinghouseState — account state for a single dex (omit dex for main perps)
getOpenOrders(user?)	OpenOrder[] — all open orders across all dexes. Each: { coin, side, limitPx, sz, oid, timestamp, orderType }
getUserFills(user?, aggregateByTime?)	Array<{ coin, px, sz, side, time, closedPnl, fee, oid, tid, crossed, builderFee }> — trade fill history. side: 'B' (buy) or 'A' (sell)
getHistoricalOrders(user?)	Array<{ order: { coin, side, limitPx, sz, origSz, oid, timestamp, orderType, tif, triggerCondition, triggerPx, isTrigger, isPositionTpsl, reduceOnly }, status, statusTimestamp }> — all orders (filled, cancelled, etc.)
getOrderStatus(oid, user?)	{ status, order? } — status of a specific order by numeric OID or string CLOID
getUserFunding(user?, startTime?, endTime?)	Array<{ time, hash, delta: { coin, usdc, szi, fundingRate } }> — funding payments received/paid
getUserFees(user?)	{ dailyUserVlm, feeSchedule, userCrossRate, userAddRate, activeReferralDiscount, activeStakingDiscount } — fee tier, rates, and volume
getUserRateLimit(user?)	{ cumVlm, nRequestsUsed, nRequestsCap, nRequestsSurplus } — API rate limit status
getSpotBalances(user?)	{ balances: Array<{ coin, token, hold, total, entryNtl }> } — spot token balances
getSubAccounts(user?)	Array<{ subAccountUser, name }> — sub-accounts for a master wallet
getAccountMode(user?)	string — account abstraction mode: 'standard', 'unified', 'portfolio', or 'dexAbstraction'
isUnifiedAccount(user?)	boolean — true if unified or portfolio margin (shared USDC across dexes)

Leverage & Config

Method	Description
updateLeverage(coin, leverage, isCross?)	Set leverage. isCross defaults to true (cross margin). HIP-3 assets are forced to isolated and clamped to their max leverage
approveBuilderFee(maxFeeRate?, builder?)	Approve builder fee (must be called from main wallet, not API wallet). Default rate: '0.1%'
getMaxBuilderFee(user?, builder?)	Check approved builder fee. Returns fee string (e.g. '0.01%') or null if not approved

Utility Properties

Property / Method	Description
getAssetIndex(coin)	Get numeric asset index for a coin (used internally for order wire)
getSzDecimals(coin)	Get size decimal precision for a coin
isHip3(coin)	Check if a coin is a HIP-3 asset
getCoinDex(coin)	Get dex name for a coin (null for main perps)
getAllAssetNames()	Get all known asset names (main + HIP-3)
getHip3AssetNames()	Get only HIP-3 asset names
invalidateMetaCache()	Force refresh of market metadata on next call

Utility Functions (api.utils)

Function	Description
roundPrice(price, szDecimals, isSpot?)	Round price to 5 significant figures (max 6 decimals perp, 8 spot)
roundSize(size, szDecimals)	Round size to asset-specific decimal precision
sleep(ms)	Promise-based delay
normalizeCoin(coin)	Normalize coin name (uppercase, trim whitespace)
formatUsd(amount)	Format number as USD string (e.g. $1,234.56)
annualizeFundingRate(hourlyRate)	Convert hourly funding rate to annualized percentage

Example: Price Breakout

// ~/.openbroker/automations/breakout.ts
export default function(api) {
  const COIN = 'ETH';
  const BREAKOUT_PCT = 2;  // 2% move triggers entry
  const SIZE = 0.5;
  let basePrice = null;

  api.onStart(async () => {
    const mids = await api.client.getAllMids();
    basePrice = parseFloat(mids[COIN]);
    api.log.info(`Watching ${COIN} from $${basePrice} for ${BREAKOUT_PCT}% breakout`);
  });

  api.on('price_change', async ({ coin, newPrice }) => {
    if (coin !== COIN || !basePrice) return;
    const totalChange = ((newPrice - basePrice) / basePrice) * 100;

    if (Math.abs(totalChange) >= BREAKOUT_PCT && !api.state.get('inPosition')) {
      const side = totalChange > 0;  // true = long, false = short
      api.log.info(`Breakout! ${totalChange.toFixed(2)}% — entering ${side ? 'long' : 'short'}`);
      await api.client.marketOrder(COIN, side, SIZE);
      api.state.set('inPosition', true);
    }
  });
}

Example: Scheduled DCA

// ~/.openbroker/automations/hourly-dca.ts
export default function(api) {
  const COIN = 'ETH';
  const USD_PER_BUY = 100;

  // Buy $100 of ETH every hour
  api.every(60 * 60 * 1000, async () => {
    const mids = await api.client.getAllMids();
    const price = parseFloat(mids[COIN]);
    const size = parseFloat(api.utils.roundSize(USD_PER_BUY / price, 4));
    await api.client.marketOrder(COIN, true, size);
    const count = (api.state.get('buyCount') || 0) + 1;
    api.state.set('buyCount', count);
    api.log.info(`DCA #${count}: bought ${size} ${COIN} at $${price}`);
  });
}

Example: Margin Guardian

// ~/.openbroker/automations/margin-guard.ts
export default function(api) {
  api.on('margin_warning', async ({ marginUsedPct, equity }) => {
    api.log.warn(`Margin at ${marginUsedPct.toFixed(1)}% — reducing positions`);

    // Close the smallest position to free margin
    const state = await api.client.getUserStateAll();
    const positions = state.assetPositions
      .filter(p => parseFloat(p.position.szi) !== 0)
      .sort((a, b) => Math.abs(parseFloat(a.position.positionValue)) - Math.abs(parseFloat(b.position.positionValue)));

    if (positions.length > 0) {
      const pos = positions[0].position;
      const size = Math.abs(parseFloat(pos.szi));
      const isBuy = parseFloat(pos.szi) < 0; // Close short = buy, close long = sell
      api.log.info(`Closing smallest position: ${pos.coin} (${pos.szi})`);
      await api.client.marketOrder(pos.coin, isBuy, size);
    }
  });
}

Publishing to the Agent (Webhooks)

Use api.publish() to send messages back to the OpenClaw agent. This triggers an agent turn — the agent receives the message and can notify the user via their preferred channel, take trading actions, or log the event.

// Simple notification
await api.publish(`ETH broke above $4000 — current price: $${price}`);

// With options
await api.publish(`Margin at ${pct}% — positions at risk`, {
  name: 'margin-alert',           // appears in logs
  wakeMode: 'now',                // 'now' (default) or 'next-heartbeat'
  channel: 'slack',               // target channel (optional)
});

api.publish() returns true if delivered, false if webhooks are not configured (no hooks token). It requires OPENCLAW_HOOKS_TOKEN to be set (automatically configured when running as an OpenClaw plugin).

Example: Price alert automation with publish

// ~/.openbroker/automations/price-alert.ts
export default function(api) {
  const COIN = 'ETH';
  const THRESHOLD = 4000;

  api.on('price_change', async ({ coin, newPrice, changePct }) => {
    if (coin !== COIN) return;

    const crossed = api.state.get<boolean>('crossed', false);
    if (!crossed && newPrice >= THRESHOLD) {
      api.state.set('crossed', true);
      await api.publish(
        `${COIN} crossed above $${THRESHOLD}! Price: $${newPrice.toFixed(2)} (+${changePct.toFixed(2)}%)`,
      );
    } else if (crossed && newPrice < THRESHOLD) {
      api.state.set('crossed', false);
    }
  });
}

Running Automations

CLI:

openbroker auto run my-strategy --dry       # Test without trading
openbroker auto run ./funding-scalp.ts      # Run from path
openbroker auto run my-strategy --poll 5000 # Poll every 5s
openbroker auto run --example dca --set coin=HYPE --set amount=50 --dry  # Run bundled example
openbroker auto examples                    # List bundled examples with config
openbroker auto list                        # Show available scripts
openbroker auto status                      # Show running automations

Plugin tools (for OpenClaw agents):

• ob_auto_run — { "script": "funding-scalp", "dry": true } — start an automation
• ob_auto_run — { "example": "dca", "config": { "coin": "HYPE", "amount": 50 }, "dry": true } — run a bundled example
• ob_auto_stop — { "id": "funding-scalp" } — stop a running automation
• ob_auto_list — {} — list available automations, bundled examples with config schemas, and running automations

Options:

Flag	Description	Default
--dry	Intercept write methods — no real trades	false
--verbose	Show debug output	false
--id <name>	Custom automation ID	filename
--poll <ms>	Poll interval in milliseconds	10000
--example <name>	Run a bundled example automation	-
--set key=value	Set config values (repeatable)	-

Guidelines for agents writing automations:

Risk & Safety (mandatory):

• Always attach a liquidation monitoring automation to every open position. Subscribe to margin_warning and pnl_threshold events so the user is never blindsided by liquidation risk. If no margin/liquidation automation is already running, create one before placing trades.
• Use api.publish() to notify the user of important events — position opens/closes, TP/SL triggers, large PnL swings, margin warnings, errors, and any situation that requires human attention. Do NOT silently handle critical events.
• Always register an api.onStop() handler to clean up — cancel open orders and close positions (or at minimum alert the user) on shutdown. Never leave orphaned orders or unmanaged positions.
• Do NOT use --dry unless the user explicitly asks for it. Automations should run live by default.
• Never place trades without validating that sufficient margin is available. Check account state before sizing orders.
• Cap position sizes relative to account equity. Do not risk more than a reasonable percentage of equity on a single trade unless the user explicitly specifies the size.
• Always set TP/SL on new positions — either within the automation or by confirming the user has them set. Unprotected positions are a liability.

State & Reliability:

• Use api.state to track position state, entry prices, and flags across restarts. Never rely on in-memory variables alone — automations persist across gateway restarts and are automatically restarted.
• Use idempotency guards (api.state.get/set) to prevent duplicate orders. Events can fire multiple times for the same condition across polls — always check state before placing orders.
• The runtime catches errors per handler — one failing handler won't crash others, but always handle expected errors (e.g. order rejection, insufficient margin) gracefully within handlers.

Communication:

• Use api.publish() to send alerts/events back to the OpenClaw agent — do NOT manually construct webhook requests.
• Publish on: position opened/closed, TP/SL triggered, PnL threshold exceeded, margin warning, automation errors, and any automated trade execution. The user should always know what the automation did and why.
• Include actionable context in publish messages — coin, price, size, PnL, and what happened — so the user can make informed decisions without checking the terminal.

General:

• Scripts are loaded from ~/.openbroker/automations/ by name, or from any absolute path.
• All trading commands support HIP-3 assets (api.client.marketOrder('xyz:CL', true, 1)).
• Automations persist across gateway restarts — they are automatically restarted when the gateway comes back up.
• Prefer api.every(ms, fn) over tick for periodic tasks with intervals longer than the poll cycle.

Risk Warning

• Always use --dry first to preview orders
• Start with small sizes on testnet (HYPERLIQUID_NETWORK=testnet)
• Monitor positions and liquidation prices
• Use --reduce for closing positions only