Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Query Polymarket prediction markets — search events, check odds and prices, view trending markets, track price momentum, get orderbook depth, analyze volume, and monitor market resolution timelines. Use when the user asks about prediction markets, betting odds, event probabilities, or Polymarket data.
.claude/skills/cowork-os-polymarket/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 231% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 192% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 172% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 115% | 0% |
Query Polymarket prediction markets — search events, check odds and prices, view trending markets, track price momentum, get orderbook depth, analyze volume, and monitor market resolution timelines. Use when the user asks about prediction markets, betting odds, event probabilities, or Polymarket data.
| Name | Type | Required | Description | |---|---|---|---| | query | string | No | What to search for or ask about (e.g., 'Trump election odds', 'trending crypto markets', 'AI predictions') |
../polymarket.json.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 10,835 | 9,252 | -15% | 1 | 1 | 0% | 1,676 | 1,777 | +6% | 0 | 0 | — |
case-02 | fail→pass | 6,828 | 26,003 | +281% | 1 | 1 | 0% | 1,019 | 3,374 | +231% | 0 | 0 | — |
case-03 | pass→pass | 10,655 | 16,412 | +54% | 1 | 1 | 0% | 1,783 | 2,389 | +34% | 0 | 0 | — |
case-04 | pass→fail | 2,544 | 3,313 | +30% | 1 | 1 | 0% | 393 | 843 | +115% | 0 | 0 | — |
case-05 | fail→fail | 2,649 | 3,404 | +29% | 1 | 1 | 0% | 325 | 836 | +157% | 0 | 0 | — |
case-06 | pass→pass | 3,429 | 2,686 | -22% | 1 | 1 | 0% | 522 | 708 | +36% | 0 | 0 | — |
case-07 | pass→fail | 11,984 | 5,695 | -52% | 1 | 1 | 0% | 1,299 | 635 | -51% | 0 | 0 | — |
case-08 | pass→fail | 18,464 | 1,759 | -90% | 1 | 1 | 0% | 3,203 | 565 | -82% | 0 | 0 | — |
case-18 | fail→fail | 7,636 | 5,924 | -22% | 1 | 1 | 0% | 448 | 706 | +58% | 0 | 0 | — |
case-09 | pass→fail | 17,090 | 5,471 | -68% | 1 | 1 | 0% | 1,907 | 583 | -69% | 0 | 0 | — |
case-10 | pass→fail | 14,069 | 2,746 | -80% | 1 | 1 | 0% | 2,127 | 622 | -71% | 0 | 0 | — |
case-11 | pass→pass | 14,561 | 12,186 | -16% | 1 | 1 | 0% | 2,351 | 1,941 | -17% | 0 | 0 | — |
case-12 | pass→fail | 11,429 | 6,249 | -45% | 1 | 1 | 0% | 1,913 | 774 | -60% | 0 | 0 | — |
case-13 | pass→pass | 14,146 | 16,270 | +15% | 1 | 1 | 0% | 2,209 | 1,912 | -13% | 0 | 0 | — |
case-14 | pass→pass | 9,808 | 10,193 | +4% | 1 | 1 | 0% | 1,459 | 1,572 | +8% | 0 | 0 | — |
case-15 | fail→fail | 6,827 | 5,555 | -19% | 1 | 1 | 0% | 279 | 838 | +200% | 0 | 0 | — |
case-16 | pass→fail | 15,527 | 11,455 | -26% | 1 | 1 | 0% | 2,461 | 1,404 | -43% | 0 | 0 | — |
case-17 | fail→fail | 12,274 | 6,615 | -46% | 1 | 1 | 0% | 1,738 | 1,403 | -19% | 0 | 0 | — |
case-19 | fail→pass | 4,371 | 11,283 | +158% | 1 | 1 | 0% | 534 | 1,561 | +192% | 0 | 0 | — |
case-20 | fail→pass | 8,316 | 4,705 | -43% | 1 | 1 | 0% | 388 | 1,057 | +172% | 0 | 0 | — |
case-21 | pass→pass | 9,955 | 7,954 | -20% | 1 | 1 | 0% | 1,528 | 1,606 | +5% | 0 | 0 | — |
case-22 | pass→pass | 18,394 | 9,496 | -48% | 1 | 1 | 0% | 2,452 | 2,010 | -18% | 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. 22 cases were attempted, and 16 counted toward the lift figure. The other 6 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 -14 percentage points is the difference between those two pass rates over the 16 comparable cases. 7 cases got worse with the skill loaded, and they are included in that figure.
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.