machina-sports/sports-skillsMIT6 files

Football Data

|

Specification
Skill ID
machina-sports/sports-skills/football-data
Publisher
machina-sports
Repository
sports-skills
Installs
0
Files
6
License
MIT
Synced
Sep 23, 2026
How to use it

Open any RiverX project, open the Skills panel in the chat, and search for this identifier. The files are fetched from the source repository at install time.

machina-sports/sports-skills/football-dataInstalls these files
  • SKILL.md
  • references/api-reference.md
  • references/commands.md
  • references/data-coverage.md
  • references/schemas.md
  • scripts/validate_params.sh

What this skill tells the agent

Football Data

Before writing queries, consult references/api-reference.md for endpoints, ID conventions, and data shapes.

Setup

Before first use, check if the CLI is available:

which sports-skills || pip install sports-skills

If pip install fails (package not found or Python version error), install from GitHub:

pip install git+https://github.com/machina-sports/sports-skills.git

The package requires Python 3.10+. If your default Python is older, use a specific version:

python3 --version  # check version
# If < 3.10, try: python3.12 -m pip install sports-skills
# On macOS with Homebrew: /opt/homebrew/bin/python3.12 -m pip install sports-skills

No API keys required.

Quick Start

Prefer the CLI — it avoids Python import path issues:

sports-skills football get_daily_schedule
sports-skills football get_season_standings --season_id=premier-league-2025

Python SDK (alternative):

from sports_skills import football

standings = football.get_season_standings(season_id="premier-league-2025")
schedule = football.get_daily_schedule()

CRITICAL: Before Any Query

CRITICAL: Before calling any data endpoint, verify:

  • Season ID is derived from get_current_season(competition_id="...") — never hardcoded.
  • Team ID is resolved via search_team(query="...") and passed as the numeric team_id. For get_head_to_head, get_team_strength, and get_match_forecast, always pass IDs — ambiguous names (e.g. two "Paris" clubs) can resolve to the wrong team.
  • The endpoint actually covers the league in question — see the Coverage & Source Map below. Coverage is uneven across sources; an uncovered call returns an empty payload with a message, not data.
  • get_event_xg and get_event_players_statistics (with xG) are only called for top-5 leagues (EPL, La Liga, Bundesliga, Serie A, Ligue 1).
  • get_season_leaders and get_missing_players are only called for Premier League seasons (season_id must start with premier-league-).

Choosing the Season

Derive the current year from the system prompt's date (e.g., currentDate: 2026-02-16 → current year is 2026).

  • If the user specifies a season, use it as-is.
  • If the user says "current", "latest", or doesn't specify: Call get_current_season(competition_id="...") to get the active season_id. Do NOT guess or hardcode the year.
  • Season format: Always {league-slug}-{year} (e.g., "premier-league-2025" for the 2025-26 season). The year is the start year of the season, not the end year.
  • MLS exception: MLS runs spring-fall within a single calendar year. Use get_current_season(competition_id="mls").

Coverage & Source Map

This skill stitches several free sources together. Coverage is not uniform — each endpoint works only where its underlying source has data. Check this before promising an answer; when an endpoint isn't covered, it returns an empty payload with an explanatory message (never an error) — read that message and fall back.

Endpoint(s)SourceCoverage
standings, schedules, teams, event summary/lineups/stats/timelineESPNAll leagues (broadest — the backbone)
get_event_xg, get_event_players_statistics (xG fields)UnderstatTop 5 only (EPL, La Liga, Bundesliga, Serie A, Ligue 1). Not RFPL — Understat dropped it.
get_season_leaders, get_missing_playersFPLPremier League only
get_player_profile, get_season_transfers (market value)TransfermarktAny player with a tm_player_id
get_head_to_headfootball-data.co.uk11 European domestic leagues (EPL, Championship, La Liga, Serie A, Bundesliga, Ligue 1, Eredivisie, Primeira Liga, Scottish, Belgian, Turkish). Same-division meetings only.
get_team_strengthClubElo, falling back to local EloEuropean clubs (incl. Russia). Falls back to ratings computed from football-data.co.uk when ClubElo is down.
get_match_forecastClubEloEuropean clubs (incl. Russia). No fallback — needs ClubElo's fixture feed.

Rule of thumb: ESPN answers "what happened" everywhere; the enrichment sources ( Understat/FPL/ClubElo/football-data.co.uk ) add depth only in their coverage zone. ESPN is always the fixture/score authority — never let an enrichment source override an ESPN score.

Gotchas (from live testing)

  • `get_team_profile` returns the squad. data.players[] carries the current roster with ESPN athlete ids, shirt numbers and ages — use it instead of collecting names match by match.
  • `get_player_season_stats` takes the same league slug as everything else (serie-a-brazil, not only ESPN's bra.1), and its gamelog is the last ~5 matches across competitions, not a season total.
  • Scored penalties are `penalty_goal` in the timeline. Count goal + penalty_goal + own_goal when reconciling with the score.
  • Pass IDs, not ambiguous names. For H2H/strength/forecast, resolve teams with search_team first and pass the numeric team_id. Names like "Paris Saint-Germain" can collapse onto the wrong club (Paris FC) during name resolution.
  • ClubElo off-season gaps: current-date get_team_strength can miss clubs in the summer break (a club's weekly Elo period may not span today). If a well-known club returns unresolved, pass an in-season date (e.g. date="2026-03-01").
  • ClubElo outages: get_team_strength falls back to locally computed Elo and sets source: "local-elo". Check that field before comparing numbers across calls — the local scale is division-local, so a rating means nothing outside its own division and cross-division comparisons are refused. The fallback honours date (it rates the division as of that date, and each entry's as_of is the last match counted). get_match_forecast has no fallback and stays empty.
  • `get_match_forecast` is short-horizon: ClubElo only forecasts ~a week ahead — empty between matchdays / off-season. That's expected, not a failure.
  • H2H is same-division only: two clubs that met in a cup or across tiers won't show; it counts league meetings in the resolved division.
  • H2H tells "unresolved" apart from "never met": football-data.co.uk uses short exonyms/abbreviations ("FC Koln", "M'gladbach", "Sp Lisbon"). Each club in teams[] reports resolved + matched_as; if a club is resolved: false, zero meetings means the lookup failed, not that the clubs never played.

Combining Endpoints (mix-and-match)

Compose sources for richer answers. Run independent calls in parallel.

  • Match preview (X vs Y): search_team ×2 → get_head_to_head (recent record) + get_team_strength(team_id, team_id_2) (Elo gap / favorite) + get_match_forecast (if within ~a week: W/D/L + scoreline). For a top-5 fixture add historical get_event_xg context from recent meetings.
  • Match report (post-game): get_event_summary + get_event_statistics + get_event_timeline, and for top-5 leagues get_event_xg + get_event_players_statistics.
  • Team form + context: get_team_schedule (recent results) + get_team_strength (current Elo & rank) + get_missing_players (PL only) + per-match get_event_xg (top-5).
  • Rivalry / derby deep dive: get_head_to_head (all-time-ish record + goals) + get_team_strength comparison for the current power balance.
  • Odds sanity-check: get_match_forecast gives a free model baseline (W/D/L) to compare against the kalshi / polymarket betting skills.

When a piece of the composition isn't covered (e.g. xG outside the top 5, H2H for MLS), skip it silently and deliver the parts that are covered — don't block the whole answer on one missing source.

Commands

CommandDescription
get_current_seasonDetect current season for a competition
get_competitionsList available competitions with current season info
get_competition_seasonsAvailable seasons for a competition

Use this skill in a project

Open any RiverX project, search the Skills panel for machina-sports/sports-skills/football-data, and install it in one click.

Use this skill