superior-trade/superior-skills/polymarket
Claude Code and agent skill by superior-trade/superior-skills
- community
Checking that the SKILL.md file is still in the repository.
Use when the user wants to trade, research, or backtest Polymarket prediction markets through Superior Trade — finding markets by slug or event URL, placing a single immediate market order, writing NautilusTrader strategies, running filled-data backtests, funding pUSD, or deploying and monitoring a live Polymarket strategy.
Install
npx vskill@latest install superior-trade/superior-skills/polymarketSKILL.md
Polymarket Prediction Market Trading
Trade prediction markets on Polymarket through Superior Trade. Discover markets, place single immediate market orders, write NautilusTrader strategies, backtest against historical trade data, and deploy live — all through one API.
Read `../../references/unified-runtime.md` before any Superior Trade request. Use only methods, fields, and capabilities published by Unified OpenAPI.
Base URL: https://api.superior.trade. If the environment configures a different base URL, use that instead. Auth: Prefer x-api-key: <api_key> for Superior Trade product API keys. Browser/session callers may use Authorization: Bearer <token>. Docs: GET /openapi.json (Unified OpenAPI contract)
Reference files
Load these on demand — each is the full detail behind a summary below.
| Read | When | |
|---|---|---|
references/api.md | You need Unified route selection for account, wallet, market discovery, typed execution, backtest, or deployment. | |
references/strategy-authoring.md | You are writing NautilusTrader strategy code — structure, order submission, dollars-to-shares conversion, multi-instrument setups, available data. | |
references/example-strategies.md | You want a worked starting point: carry/yield harvesting, momentum/news fade, or spread capture. | |
references/strategies.md | You are choosing an archetype for the user's thesis and need the fit/anti-fit table. | |
references/troubleshooting.md | A strategy source, backtest, or deployment is failing, or a deployment trades zero times. |
Gotchas
Environment-specific facts that defy reasonable assumptions. Read these before acting.
- Wallet reads are not bootstrap or funding mutations. Use
GET /walletonly for the managed wallet data it returns. If a required Polymarket funding action is absent from the Unified contract, report it as unavailable. - Backtests are built from filled `TradeTick` data, not a full order book. They cannot model queue position or liquidity, so a strategy that depends on resting-order fills will look better in backtest than it trades. Say this when presenting results.
- An exact market slug is required before writing a strategy. Get it from
GET /context/markets; if the user pastes a Polymarket event URL, pass the whole URL as the query so child markets expand. - Prices are probabilities between 0 and 1. Sizing is in shares, not dollars — convert with
shares = dollars / price(seereferences/strategy-authoring.md). - Credential storage is contract-defined. Use only the form published for
PUT /runtime/deployments/{id}/credentials; never infer or send undisclosed key material.
Setup
API Key
This skill uses the same Superior Trade API key as every other venue — there is no separate prediction-market key:
If SUPERIOR_TRADE_API_KEY is not set, use the registration and verification flow in the superior-trade-auth skill.
Wallet and Funding
The managed wallet is read through /wallet. Polymarket support is discovered through /context/venues; exact capabilities come from /openapi.json.
Before live trading:
- Read the managed wallet with
GET /wallet. - Check Polymarket and Nautilus support with
GET /context/venuesandGET /runtime/frameworks. - Confirm the wallet has the required balance and readiness exposed by Unified API.
- If a required onboarding or pUSD funding action is absent from
GET /openapi.json, stop and report it. - Create a Unified deployment with strategy code and config.
- Store credentials with
PUT /runtime/deployments/{id}/credentialsonly when the current contract requires it. - Start the deployment only after the user explicitly confirms.
Use only the credential schema published by Unified API. Never infer a secret or wallet field from an older payload.
Account Endpoints
Current account/funding endpoints:
GET /accountGET /walletGET /wallet/depositsGET /wallet/withdrawalsPOST /wallet/withdrawGET /context/venuesPOST /runtime/executions
Single Immediate Market Orders
Use POST /runtime/executions when the user asks to place one immediate Polymarket bet/order. This is a fast path for a whitelisted placeMarketOrder action, not a strategy deployment. Do not create a deployment plan or run a backtest for a one-off order unless the user asks for a strategy or bot.
Required sequence:
- Confirm auth with
SUPERIOR_TRADE_API_KEY. - Fetch
GET /wallet. - If selecting a non-default managed account, set top-level
account_addressonly to an address verified byGET /wallet. - Check
GET /context/venuesand the wallet balance. If the required capability or collateral is unavailable, stop and report the blocker. - If the user already provided the complete order payload and explicit affirmative confirmation for that exact payload in the current conversation, submit it without asking unrelated follow-up questions. If not, show the exact payload and ask for confirmation before submitting.
- Send exactly one request to
POST /runtime/executionsand report the real API result. Never fabricate fills, order IDs, balances, or status.
Supported action:
Fast confirmation format:
Inspect the durable execution with GET /runtime/executions/{id}. If the Unified contract does not expose the current position or open-order read the user needs, say that it is unavailable instead of querying with undisclosed venue credentials.
Safety
Security & Permissions
This skill requires exactly one credential: a Superior Trade API key. The only secret the agent uses is SUPERIOR_TRADE_API_KEY.
Security rules (non-negotiable):
- NEVER ask users for private keys, seed phrases, or wallet credentials
- NEVER log, store, or display private keys or seed phrases
- NEVER fabricate wallet balances, API responses, market prices, or trade results
- NEVER start a live deployment without explicit user confirmation
- NEVER promise a venue withdrawal or portfolio exit unless the exact operation is present in Unified OpenAPI
- Prefer user-friendly language over internal technical names when speaking conversationally. Say "strategy" or "the bot" instead of internal class names or infrastructure details. If the user asks about the underlying technology, answer honestly (the platform uses NautilusTrader for strategy execution against Polymarket's CLOB).
| Can do | Cannot do | |
|---|---|---|
| Read Unified account, wallet, venue, and market context | Access other users' data | |
| Run supported filled-data backtests | Export, accept, or view private keys | |
| Create and manage Unified deployments | Guess holdings from deployment state | |
| Place or cancel a typed Polymarket order after confirmation | Invent balances or trade results | |
| Inspect durable execution records | Transfer or bridge arbitrary funds |
Polymarket-Specific Risks (tell the user when relevant)
- Prices are probabilities — every outcome trades between 0.001 and 0.999 pUSD. A "cheap" 0.03 outcome is not a bargain; it is a ~3% market-implied probability.
- Resolution risk — positions held to resolution settle at exactly 0 or 1. A strategy can be up and still settle at zero if the outcome resolves against it.
- Liquidity risk — many markets have thin order books. Market orders can fill far from the displayed price. Prefer markets with meaningful
volume_24handliquidity. - Market end dates — markets stop trading at
end_date. Never deploy a strategy on a market that resolves before the strategy has time to work. - Rate limits — Polymarket enforces ~30 order submissions/minute and ~100 data requests/minute per user. Strategies that cancel/replace on every tick can hit these fast.
Live Deployment Confirmation
Before any live deployment start, the agent MUST present this summary and wait for an explicit affirmative response:
Do NOT start a live deployment without an explicit affirmative response.
Agent Operating Rules
- Verification-first: Every factual claim about balance, market price, position, or deployment status MUST be backed by an API call in the current turn. NEVER assume → report → verify later.
- Anti-hallucination: If you can't call the API, say "I haven't checked yet." Every number must come from a real response.
- Position/order state: When the user asks what they hold, whether orders are open, whether a bot traded, or whether it is safe to stop/exit, first check Unified OpenAPI for a supported live position/order read. If none exists, say exposure verification is unavailable; do not infer it from deployment status or proceed with an exit, replacement, restart, or exposure-based sizing decision whose safety depends on that state.
- Backtest before deploy: Always run a backtest and review results with the user before the first live deployment of any strategy.
- Conversational: Make API calls directly and present results conversationally. Show raw payloads only on request.
- Proactive: Ask for missing info conversationally, one concern at a time.
Repeated Failures
If the same task fails 3+ times (e.g. strategy source/config keeps failing, backtest keeps erroring), stop and:
- Summarize what was tried and what failed
- Suggest a simpler approach or different parameters
- If the issue appears to be model capability (complex multi-instrument logic), suggest switching to a more capable model for strategy generation
Workflow
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/polymarket/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/superior-trade/superior-skills/polymarket)