machina-sports/sports-skills
5 skills3.4k installs670 starsGitHub
Install
npx skills add https://github.com/machina-sports/sports-skillsSkills in this repo
1Football DataFootball Data is an agent skill that documents how to query multi-source football datasets with predictable identifiers and clear league coverage boundaries. Indie builders shipping fantasy helpers, match preview apps, or internal analytics copilots install it so agents call the right commands for ESPN-wide fixtures versus Understat xG-only top leagues versus FPL-exclusive Premier League signals. The readme maps season and competition slugs, numeric ESPN team and event IDs, FPL element codes, and Transfermarkt player IDs so prompts do not mix incompatible keys across providers. It emphasizes what each data source actually provides—scores and lineups everywhere, expected goals where Understat applies, injury and ownership depth only for PL through FPL—reducing wasted agent loops on unsupported leagues. Use during backend integration and agent-tooling design when your product needs live football context without building scrapers from scratch.1.1kinstalls2PolymarketPolymarket is an agent skill packaged as an API reference for sports prediction markets on Polymarket. Solo builders shipping dashboards, alerts, or trading assistants use it so Claude Code or similar agents call the right command names with the correct required fields—without re-reading scattered docs. The core command set covers discovery and pricing: league configs, today’s events, market search, series listings, single market and event details, live CLOB prices, full order books, and historical price series. A separate trading section documents limit and market orders plus wallet configuration, clearly separated from read-only calls that need no extra dependencies. Parameter tables distinguish required versus optional inputs (sport, slug, token_id, game_id, sports_market_types, etc.), which reduces bad tool invocations during iterative agent builds. It is narrowly integration-focused: it does not replace product strategy, compliance review, or bankroll logic, but it accelerates backend wiring and agent tooling for sports-market automation.664installs3Bettingbetting is an agent skill that packages a sports wagering analysis API for solo builders shipping odds comparison tools, fantasy edges, or research agents. It is reference-first: tables for American, decimal, and probability formats, conversion rules including Kalshi integer prices, and a matrix of commands from simple convert_odds through devig, find_edge, kelly_criterion, evaluate_bet, find_arbitrage, parlay_analysis, and line_movement. If you are one person building a SaaS or CLI that must speak US books and European decimals—or blend sportsbooks with prediction markets—this skill keeps probability math consistent instead of re-deriving vig removal and Kelly sizing in every prompt. It fits the Build integrations shelf because you invoke named commands while wiring backends or agent tools, not while doing unrelated frontend polish. Complexity is intermediate: you need comfort with implied probability and expected value, but the skill states the formulas explicitly.559installs4Sports NewsSports News is an agent skill that documents how to pull sports headlines through RSS/Atom feeds and Google News search. Indie builders shipping fantasy assistants, club apps, betting-adjacent content sites, or Telegram digests use it so agents respect constraints: Google News mode always needs a query, feed URL mode cannot mix with Google News, and recency tasks should set `sort_by_date=True` with `after` derived from the session’s current date. The reference defines `fetch_feed` for single-feed introspection and `fetch_items` for bounded lists with optional limits and date windows. A separate curated URL list points to major publishers, which speeds up prototyping without hunting feeds manually. During Grow you wire this into content pipelines and notification copy; during Build integrations you embed the same commands behind your API layer. The skill teaches command contracts and feed sources rather than shipping a hosted news API key.552installs5Nba DataNBA Data is an agent skill that documents Machina Sports NBA API commands for solo builders shipping fantasy tools, betting-adjacent dashboards, fan apps, or editorial sports products. It maps each operation—get_scoreboard, get_standings, get_teams, get_team_roster, get_team_schedule, get_game_summary, get_leaders, and get_news—to parameters and response shapes (events, competitors, articles) so your coding agent can call the right endpoint with correct IDs and dates. Use it when you need authoritative league context during feature work rather than guessing field names or duplicating scrape logic. It does not replace product design or compliance review for gambling or minors-facing apps; pair it with your own caching, rate limits, and display rules. The reference is task-focused integration knowledge, not a full sports analytics pipeline.513installs