Polymarket Agent Skills

Overview

Polymarket Agent Skills is an official repository published by Polymarket on GitHub (Polymarket/agent-skills), last updated February 19, 2026. It provides a structured set of documentation files designed to be consumed by AI agents, coding assistants, or agent frameworks as context for interacting with the Polymarket prediction market API.

Unlike executable SDKs (py-clob-client, @polymarket/clob-client), Agent Skills is a knowledge layer. Each skill file covers a specific capability with explanations, code patterns, and API reference details in a format optimized for LLM comprehension.

Skill Coverage

Skill Area	What It Covers
Authentication	EIP-712 signatures, wallet types, signature_type (0/1/2), API credential derivation. Emphasizes signature_type=2 (Gnosis Safe) as the most common wallet type.
Order Patterns	Limit orders, market orders, batch orders (up to 15), post-only orders. OrderArgs vs. MarketOrderArgs.
Market Data	Gamma API for discovery, CLOB API for order books and prices, Data API for positions and history. Token ID lookups and neg-risk market handling.
WebSocket Streaming	Market, User, Sports, and RTDS channels. Subscription formats, heartbeat intervals, orderbook reconstruction.
CTF Operations	Conditional Token Framework: split, merge, and redeem operations for outcome tokens. On-chain settlement flow.
Bridge	Deposit and withdrawal flows via the Bridge API, including the /withdraw endpoint for cross-chain bridging of USDC.e.
Gasless Transactions	Builder Program Relayer Client for submitting orders without end-user gas costs. HMAC-signed builder headers and BuilderConfig setup.

Use Cases

Standalone agent context — Feed skill files directly to an LLM-based agent as system context so it understands how to call Polymarket APIs. Works with any LLM framework (LangChain, CrewAI, custom loops).

OpenClaw integration — Mount the skill files as reference documents in an OpenClaw agent configuration. The agent can then autonomously place trades, monitor markets, and manage positions on Polymarket.

Developer reference — Even for human developers, the skill files serve as a concise reference that covers common patterns without the noise of full API documentation.

Quick Start

git clone https://github.com/Polymarket/agent-skills.git

Then point your agent framework at the cloned directory. For example, with a simple Python agent:

from pathlib import Path

skills_dir = Path("agent-skills/skills")
auth_skill = (skills_dir / "authentication.md").read_text()
order_skill = (skills_dir / "order-patterns.md").read_text()

system_prompt = f"""You are a Polymarket trading agent.

{auth_skill}

{order_skill}
"""

Comparison

Tool	Type	Purpose
agent-skills	Documentation / skill pack	Teach AI agents how to use Polymarket APIs
py-clob-client	Python SDK	Execute API calls programmatically
@polymarket/clob-client	TypeScript SDK	Execute API calls in TypeScript/Node
polymarket-apis	Community Python wrapper	Unified wrapper around all Polymarket APIs
Polymarket CLI	Command-line tool	Terminal-based market access with JSON output

Links

GitHub: Polymarket/agent-skills
Polymarket API Guide: agentbets.ai/guides/polymarket-api-guide/
py_clob_client Reference: agentbets.ai/guides/py-clob-client-reference/
OpenClaw Guide: agentbets.ai/guides/openclaw-prediction-market-guide/
Polymarket Changelog: docs.polymarket.com/changelog/changelog

Frequently Asked Questions

What is Polymarket Agent Skills?

It is an official GitHub repository (Polymarket/agent-skills) containing structured documentation files designed to be consumed by AI agents. Each file covers a specific capability — authentication, order placement, market data, WebSocket streaming, CTF operations, bridging, and gasless transactions — so agents can learn how to interact with Polymarket's APIs.

How do I use Polymarket Agent Skills with my AI agent?

Clone or download the repo, then feed the relevant skill files as context to your LLM-based agent. If you use an agent framework like OpenClaw, you can mount the skill files as reference documents. The skills are plain text/markdown designed for LLM consumption, not executable code.

Does Polymarket Agent Skills include code examples?

Yes. The skill files include Python and TypeScript code snippets for each capability, using the official py-clob-client and @polymarket/clob-client SDKs. These are illustrative examples within the documentation, not a standalone runnable package.

Is Polymarket Agent Skills the same as py-clob-client?

No. py-clob-client is an executable Python SDK for making API calls. Agent Skills is a documentation-first resource that teaches AI agents how to use py-clob-client and the broader Polymarket API ecosystem. They are complementary — Agent Skills provides the knowledge, py-clob-client provides the implementation.