Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Get live scores, schedules, and standings for major leagues with zero API keys — ESPN's public JSON endpoints via curl, covering NFL, NBA, MLB, NHL, and world football. Use when asked what's the score, did my team win, today's games, or league standings right now. Produces the scores with game state (live/final/scheduled), the asked-team answer first, and the rerunnable command — with the unofficial-API caveat stated.
.claude/skills/mohitagw15856-sports-scores/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 622% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 127% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 395% | 0% |
"Did we win?" is a live-data question with a keyless answer: ESPN serves structured JSON scoreboards for every major league over plain HTTPS — no key, no signup. The endpoints are unofficial (publicly accessible but undocumented, and they occasionally reshape), so this skill pairs them with the discipline that makes unofficial sources usable: state the source, expect drift, and never present a stale scoreboard as live.
Ask for these if not provided:
curl -s "https://site.api.espn.com/apis/site/v2/sports/{sport}/{league}/scoreboard" — pairs: football/nfl, basketball/nba, baseball/mlb, hockey/nhl, basketball/wnba, football/college-football, soccer/eng.1 (Premier League; other comps: esp.1, ita.1, ger.1, uefa.champions, usa.1). Add ?dates=YYYYMMDD for a specific day. Each event: competitions[0].competitors[] (team.displayName, score, homeAway), status.type (state: pre/in/post, detail: "Final", "End of 3rd")..../{sport}/{league}/teams for the id map · .../{sport}/{league}/standings where available · a team's schedule: .../teams/{id}/schedule.The direct answer: "49ers won 27–20 (final, last night)" or "Down 95–98, 4:12 left in the 4th — as of 21:42."]
| Matchup | Score | State | Time (local) | |---|---|---|---| The slate, when a league day was asked]
Source: ESPN public feed (unofficial) · as of fetch time] · rerun: [exact curl]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 15,552 | 12,179 | -22% | 1 | 1 | 0% | 1,646 | 2,332 | +42% | 0 | 0 | — |
case-02 | fail→fail | 25,195 | 8,219 | -67% | 1 | 1 | 0% | 939 | 2,460 | +162% | 0 | 0 | — |
case-03 | fail→pass | 16,270 | 14,369 | -12% | 1 | 1 | 0% | 377 | 2,721 | +622% | 0 | 0 | — |
case-04 | pass→pass | 25,599 | 25,788 | +1% | 1 | 1 | 0% | 3,002 | 3,604 | +20% | 0 | 0 | — |
case-05 | fail→fail | 5,357 | 10,424 | +95% | 1 | 1 | 0% | 820 | 1,891 | +131% | 0 | 0 | — |
case-06 | pass→pass | 16,559 | 16,616 | +0% | 1 | 1 | 0% | 2,208 | 2,811 | +27% | 0 | 0 | — |
case-07 | fail→fail | 13,227 | 16,355 | +24% | 1 | 1 | 0% | 1,344 | 2,651 | +97% | 0 | 0 | — |
case-08 | fail→pass | 5,608 | 6,518 | +16% | 1 | 1 | 0% | 947 | 2,153 | +127% | 0 | 0 | — |
case-09 | fail→fail | 8,819 | 14,051 | +59% | 1 | 1 | 0% | 372 | 2,606 | +601% | 0 | 0 | — |
case-10 | fail→fail | 7,986 | 14,597 | +83% | 1 | 1 | 0% | 466 | 2,589 | +456% | 0 | 0 | — |
case-11 | fail→pass | 21,190 | 15,201 | -28% | 1 | 1 | 0% | 3,652 | 3,373 | -8% | 0 | 0 | — |
case-12 | fail→pass | 8,506 | 11,072 | +30% | 1 | 1 | 0% | 559 | 2,766 | +395% | 0 | 0 | — |
case-13 | fail→pass | 17,384 | 12,585 | -28% | 1 | 1 | 0% | 268 | 2,215 | +726% | 0 | 0 | — |
case-14 | fail→pass | 8,890 | 10,531 | +18% | 1 | 1 | 0% | 481 | 2,479 | +415% | 0 | 0 | — |
case-15 | fail→pass | 10,096 | 12,270 | +22% | 1 | 1 | 0% | 722 | 2,201 | +205% | 0 | 0 | — |
case-16 | fail→pass | 7,928 | 12,464 | +57% | 1 | 1 | 0% | 390 | 2,195 | +463% | 0 | 0 | — |
case-17 | fail→pass | 12,454 | 8,365 | -33% | 1 | 1 | 0% | 591 | 2,356 | +299% | 0 | 0 | — |
case-18 | fail→pass | 8,561 | 19,667 | +130% | 1 | 1 | 0% | 424 | 3,520 | +730% | 0 | 0 | — |
case-19 | fail→pass | 15,122 | 15,625 | +3% | 1 | 1 | 0% | 245 | 2,901 | +1084% | 0 | 0 | — |
case-20 | fail→pass | 16,200 | 10,500 | -35% | 1 | 1 | 0% | 1,379 | 1,982 | +44% | 0 | 0 | — |
case-21 | fail→pass | 19,229 | 12,192 | -37% | 1 | 1 | 0% | 2,372 | 2,249 | -5% | 0 | 0 | — |
case-22 | fail→pass | 13,794 | 12,957 | -6% | 1 | 1 | 0% | 1,384 | 2,496 | +80% | 0 | 0 | — |
case-23 | pass→pass | 12,025 | 11,660 | -3% | 1 | 1 | 0% | 1,778 | 2,215 | +25% | 0 | 0 | — |
case-24 | fail→pass | 16,357 | 7,353 | -55% | 1 | 1 | 0% | 2,182 | 2,042 | -6% | 0 | 0 | — |
case-25 | fail→fail | 11,619 | 10,947 | -6% | 1 | 1 | 0% | 966 | 1,952 | +102% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted, and 22 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.