AgentSkillsCN

kalshi-api-integration

Kalshi API 集成文档，涵盖身份验证、市场数据获取、行情解析，以及投注下单流程。

SKILL.md
--- frontmatter
name: kalshi-api-integration
description: Documentation for integrating with the Kalshi API, covering authentication, market fetching, ticker parsing, and bet placement.
version: 1.0.0

Kalshi API Integration

Authentication

Credentials stored in kalshkey (never commit):

code
API_KEY_ID
PRIVATE_KEY_PEM_CONTENT

Load with:

python
with open("kalshkey") as f:
    api_key_id = f.readline().strip()
    private_key = f.read()

API Client Setup

python
from kalshi_python import Configuration, ApiClient, MarketsApi

config = Configuration(host="https://api.elections.kalshi.com/trade-api/v2")
config.api_key['api_key_id'] = api_key_id
config.api_key['private_key'] = private_key_pem

client = ApiClient(configuration=config)
markets_api = MarketsApi(client)

Market Ticker Format

Pattern: SERIES-YYMMDD-AWAYHOME-OUTCOME

Examples:

  • •NBAWIN-250115-LAKBOS-LAK - Lakers to beat Celtics
  • •NHLWIN-250115-TORBOS-TOR - Maple Leafs to beat Bruins

Parsing Tickers

python
parts = ticker.split('-')
# parts[0]: Series (NBAWIN, NHLWIN)
# parts[1]: Date (YYMMDD)
# parts[2]: Teams (AWAYHOME combined)
# parts[3]: Winner team code

Team Name Resolution

Use NamingResolver for consistent team names:

python
from naming_resolver import NamingResolver

resolver = NamingResolver()
canonical = resolver.resolve("LAL", "nba")  # → "Los Angeles Lakers"

Fetching Markets

python
def fetch_nba_markets():
    """Fetch open NBA markets from Kalshi."""
    response = markets_api.get_markets(
        series_ticker="NBAWIN",
        status="open",
        limit=100
    )
    return response.to_dict().get("markets", [])

Saving to Database

Always save to PostgreSQL, not JSON:

python
from db_manager import default_db

def save_to_db(sport: str, markets: list):
    for market in markets:
        # Parse market data
        # Insert into game_odds table
        default_db.execute("""
            INSERT INTO game_odds (game_id, source, yes_price, no_price, ...)
            VALUES (:game_id, 'kalshi', :yes_price, :no_price, ...)
            ON CONFLICT (game_id, source) DO UPDATE SET ...
        """, params)

Bet Placement

python
from kalshi_betting import place_bet

result = place_bet(
    ticker="NBAWIN-250115-LAKBOS-LAK",
    side="yes",  # or "no"
    amount=10.00,
    limit_price=0.65  # Max price willing to pay
)

Files to Reference

  • •plugins/kalshi_markets.py - Market fetching
  • •plugins/kalshi_betting.py - Bet placement
  • •plugins/naming_resolver.py - Team name mapping