AgentSkillsCN

simmer-arbscanner

通过 Simmer 检测并执行 Polymarket 的套利机会。采用单一条件（YES+NO≠$1）与负风险策略（多结果之和≠100%）。

SKILL.md
--- frontmatter
name: simmer-arbscanner
description: Detect and execute Polymarket arbitrage opportunities via Simmer. Single-condition (YES+NO≠$1) and NegRisk (multi-outcome sum≠100%) strategies.
metadata: {"openclaw":{"emoji":"📊","requires":{"env":["SIMMER_API_KEY"]},"cron":"*/30 * * * *"}}
authors:
  - Simmer (@simmer_markets)
attribution: "Detection logic based on runesatsdev/polymarket-arbitrage-bot and IMDEA Networks research"
version: "1.0.0"

Simmer Arbitrage Scanner

Detect and execute Polymarket arbitrage opportunities using Simmer's managed execution.

When to Use This Skill

Use when you want to:

  • •Scan for arbitrage opportunities on Polymarket
  • •Detect mispriced binary markets (YES + NO ≠ $1.00)
  • •Find NegRisk opportunities (multi-outcome markets with sum ≠ 100%)
  • •Execute arb trades through Simmer's safety rails

Strategies

1. Single-Condition Arbitrage

When YES + NO prices don't sum to $1.00, guaranteed profit exists.

Example:

code
Market: "Will it rain tomorrow?"
YES price: $0.53
NO price: $0.42
Sum: $0.95

Action: Buy both for $0.95 → Payout $1.00 → Profit $0.05 (5.3% ROI)

Historical: $10.58M extracted across 7,051 conditions (IMDEA research)

2. NegRisk Rebalancing

Multi-outcome markets (3+ options) where probabilities don't sum to 100%.

Example:

code
Market: "Who wins the election?"
Candidate A: 45%
Candidate B: 46%
Candidate C: 6%
Sum: 97%

Action: Buy all for $0.97 → Payout $1.00 → Profit $0.03 (3% ROI)

Historical: $28.99M extracted, 29× capital efficiency vs single-condition

Setup

  1. •Set your Simmer API key:
bash
export SIMMER_API_KEY="sk_live_..."
  1. •Run the scanner:
bash
python arb_scanner.py              # Scan only
python arb_scanner.py --execute    # Scan and trade
python arb_scanner.py --dry-run    # Show what would trade

Configuration

Edit thresholds in arb_scanner.py:

python
MIN_PROFIT_THRESHOLD = 0.02  # Minimum 2¢ profit
MIN_ROI_THRESHOLD = 0.01     # Minimum 1% ROI
DEFAULT_TRADE_SIZE = 5.0     # $5 per leg
MAX_TRADE_SIZE = 25.0        # $25 max per opportunity

Venue Selection

bash
# Practice with virtual $SIM (default)
python arb_scanner.py --execute --venue simmer

# Real USDC on Polymarket
python arb_scanner.py --execute --venue polymarket

Example Output

code
🔍 Simmer Arbitrage Scanner
============================================================
  Execution: SCAN ONLY
  Venue: simmer
  Min profit: $0.02
  Min ROI: 1%

📡 Fetching Polymarket markets...
  Found 100 active markets

🔎 Scanning for arbitrage...
  ✅ Single-condition: Will Trump win 2028 election?...
     ROI: 3.2% | Profit: $0.48 | Action: buy_both
  ✅ NegRisk (5 outcomes): Democratic VP Nominee?...
     ROI: 4.7% | Profit: $1.27 | Sum: 0.953

📊 Found 2 opportunities

🏆 Top Opportunities:
------------------------------------------------------------
1. [negrisk] Democratic VP Nominee?...
   ROI: 4.7% | Profit: $1.27 | Capital: $27.00
   Action: buy_all

2. [single_condition] Will Trump win 2028 election?...
   ROI: 3.2% | Profit: $0.48 | Capital: $15.00
   Action: buy_both

Cron Setup

Run every 30 minutes to catch opportunities:

bash
# Add to crontab
*/30 * * * * cd /path/to/skill && python arb_scanner.py --execute --venue simmer

Or use your agent's heartbeat system.

Limitations

  • •NegRisk execution: Currently detection-only. Multi-leg execution requires atomic transactions.
  • •Slippage: Scanner uses best ask prices; actual execution may differ.
  • •Speed: Arb opportunities close quickly. This scanner is for learning/small trades, not HFT.

Attribution

Detection logic based on:

Links