tools-only sports-skills
A lightweight, zero-config Python SDK and CLI for live sports data and prediction markets.
Built natively for AI agents, but works perfectly as a standalone Python library for developers. Wraps publicly available sports data sources and APIs into unified, deterministic commands.
Zero API keys. Zero signup. Just works.
📦 Installation
Install as a global CLI tool (recommended for agents):
uv tool install sports-skills
# or
pip install sports-skills
Base install includes all sports modules.
Install as a Python library:
uv add sports-skills
# or
pip install sports-skills
Optional extras:
pip install "sports-skills[all]"
pip install "sports-skills[dev]"
⚡ What's Included
- Football (Soccer): ESPN, Understat, FPL, Transfermarkt — 21 commands across 30 leagues
- US Sports: NFL, NBA, WNBA, NHL, MLB, College Football (CFB), College Basketball (CBB) — live scores, standings, depth charts, injuries, and leaders
- Tennis: ATP and WTA tournament scores, rankings, calendars, and player profiles
- Golf: PGA, LPGA, and DP World tour scorecards and leaderboards
- Racing: Formula 1 (via FastF1) — lap times, telemetry, and race results
- Prediction Markets: Polymarket & Kalshi live odds and order books
- News: Multi-sport news aggregators
💻 CLI Usage
The package exposes a sports-skills binary.
List all supported sports:
sports-skills --help
List commands for a specific sport:
sports-skills nfl --help
Execute a command:
sports-skills nfl get_scoreboard --date 2026-02-24
sports-skills football get_current_season --competition_id premier-league
sports-skills polymarket get_markets --query "super bowl"
sports-skills news fetch_items --query "Lando Norris" --limit 5
All CLI output is printed as strict JSON, making it perfect for AI agents (Claude, GPT, Gemini) to parse and reason over.
🐍 Python SDK Usage
You can use the exact same commands directly in your Python code:
from sports_skills import nfl, football, polymarket
# Get live NFL scores
scores = nfl.get_scoreboard(date="2026-02-24")
print(scores["data"]["events"])
# Get Premier League standings
table = football.get_season_standings(season_id="premier-league-2025")
print(table["data"]["standings"])
# Fetch live odds from Polymarket
markets = polymarket.get_markets(query="bitcoin")
print(markets["data"]["markets"])
🏗️ AI Agent Integration
sports-skills is built on the Anthropic Level-3 Agent capability spec. Every command is deterministic and automatically generates its own JSON Schema.
To extract the OpenAI/Anthropic compatible tool schema for any module:
from sports_skills import nfl
import json
# Returns a list of dicts formatted exactly like Anthropic/OpenAI tools
schema = nfl.generate_schema()
print(json.dumps(schema, indent=2))
License
MIT