Sports-Plugins

A curated marketplace of high-quality community sports plugins for Claude Code, OpenAI Codex, and AI coding assistants. Covers live scores, sabermetrics, fantasy sports, and sports data analytics.

Foundation: Built on the plugin architecture from anthropics/claude-plugins-official with MCP (Model Context Protocol) endpoints designed for dual compatibility with Claude Code and OpenAI Codex tool-use patterns.

Plugins

Core Plugins (/plugins)

Plugin	Description
live-scores	Real-time scores, schedules, and game updates for MLB, NFL, NBA, and college sports
college-baseball-sabermetrics	Advanced college baseball sabermetrics — wOBA, FIP, wRC+, barrel rates, pitch analytics
fantasy-sports	Fantasy roster optimization, matchup analysis, trade evaluation, and waiver wire picks
sports-data-explorer	Query stats, compare players, and explore data across all major sports
college-football-intel	College football intel — transfer portal, recruiting rankings, Big 12 standings
mlb-the-show	MLB The Show Diamond Dynasty card analytics — valuations, squad builder, investments

External Plugins (/external_plugins)

Plugin	Description
blaze-sports-intel	BSI integration — college baseball box scores, Texas Longhorns intel, and advanced analytics via blazesportsintel.com/mcp

Installation

Install any plugin directly via Claude Code:

/plugin install <plugin-name>@Sports-Plugins

Or browse available plugins:

/plugin > Discover

Plugin Structure

Every plugin follows the standard Claude Code plugin structure:

plugin-name/
├── .claude-plugin/
│   └── plugin.json      # Plugin metadata (required)
├── .mcp.json            # MCP server configuration (optional)
├── skills/              # Skill definitions (preferred)
│   ├── skill-name/
│   │   └── SKILL.md     # Model-invoked skill
│   └── command-name/
│       └── SKILL.md     # User-invoked slash command
└── README.md            # Documentation (required)

Dual Compatibility: Claude Code + OpenAI Codex

Plugins in this marketplace are designed for both Claude Code and OpenAI Codex:

• Claude Code — Native plugin support via .claude-plugin/ metadata and skills/ definitions
• OpenAI Codex — MCP server endpoints (.mcp.json) provide tool-use access compatible with Codex's function-calling patterns

The Model Context Protocol (MCP) serves as the bridge — any plugin with an .mcp.json endpoint works seamlessly with any MCP-compatible AI coding assistant.

Validation

node scripts/validate-plugins.mjs

Shared Library (/lib)

The lib/ directory contains portable TypeScript modules used across plugins:

Module	Description
sabermetrics.ts	Pure-function sabermetric calculations — wOBA, FIP, wRC+, BABIP, ISO, xFIP, CSW%, and more with NCAA-calibrated linear weights
scores-client.ts	ESPN public API client — fetches live scoreboards, normalises response data into typed GameScore objects
types.ts	Shared TypeScript interfaces — BattingLine, PitchingLine, GameScore, Player, TeamStanding, etc.

Sabermetrics Library

The formulas in lib/sabermetrics.ts are calibrated for NCAA Division I college baseball (BBCOR bats, smaller samples). They can also be used for MLB data by adjusting the linear-weight constants.

import { wOBA, fip, computeBattingAdvanced } from "./lib/sabermetrics.ts";

const advanced = computeBattingAdvanced(battingLine, parkFactor);
console.log(advanced.wOBA, advanced.wRCPlus, advanced.babip);

Contributing

See CONTRIBUTING.md for guidelines on submitting new sports plugins.

Related Projects

• anthropics/claude-plugins-official — The official Claude Code plugins directory
• ahump20/BSI — Blaze Sports Intel — the sports platform powering the BSI external plugin
• OpenAI Codex Use Cases — Codex agent patterns for tool use

License

MIT