Skip to main content

Overview

The LuckyLobster skill enables AI agents (OpenClaw, ClawdBot, MoltBot, etc.) to trade on Polymarket prediction markets. This document describes the skill capabilities and how agents should use them.

Install from ClawHub

One-click installation for compatible agents

View skill.md

Full skill definition file

Skill File URL

Point your agent to the skill file at:
The skill file is dynamically generated and includes:
  • OpenClaw metadata with emoji and environment requirements
  • Device authorization flow for linking agents
  • Complete API documentation with examples
  • Market data, trading, and settlement endpoints

Device Authorization Flow

LuckyLobster uses a device code flow for agents to authenticate, similar to how you’d link a TV app to your account.
1

Request Device Code

Agent calls POST /api/auth/device with agent name
2

User Visits Link

User opens https://luckylobster.io/link and enters the code
3

Poll for Token

Agent polls GET /api/auth/device/token until authorized
4

Store API Key

Agent saves the returned ll_* API key for future requests

Example Flow

Capabilities

Market Discovery

Market Data

Trading

Positions & Account

Settlements

Heartbeat

Autonomous Strategies

Search Tips

For a full snapshot of all active crypto markets with prices, spreads, and oracle data in one call:
For a single crypto market lookup:
For non-crypto markets, text search works well:
  • trump election → Political markets
  • superbowl winner → Sports markets
  • oscars → Entertainment markets
Strategies using marketQuery (e.g., “bitcoin”, “xrp”) also use slug-based discovery automatically for known crypto assets.

Order Types

Permissions

Linked agents receive standard permissions:

Rate Limits

  • Default: 100 requests/minute per API key
  • Headers included: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset

Error Handling

All endpoints return consistent error format:

Example Workflow

Next Steps

API Reference

Detailed endpoint documentation

Quickstart

Get started in 5 minutes