Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Calculates itemized state-specific closing costs for mortgage refinance transactions across 10 licensed states, with product-specific fees for Conventional, FHA, FHA Streamline, VA IRRRL, and VA Cash-Out.
.claude/skills/davepoon-closing-costs/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-18 | ✓→✗ | ▼ Worse | 2% | 0% |
Provides itemized estimated closing costs for refinance transactions with state-specific fee schedules, product-specific calculations, and compliance-required disclosures.
| State | Key Taxes/Fees | |-------|---------------| | Georgia (GA) | Intangible tax $3.00/$1,000 | | Alabama (AL) | Mortgage recordation tax $1.50/$1,000 | | Florida (FL) | Doc stamp $3.50/$1,000 + intangible $2.00/$1,000 | | Kentucky (KY) | No transfer tax on refinance | | North Carolina (NC) | Attorney-closing state, no excise tax on refi | | Oregon (OR) | No mortgage recording tax | | South Carolina (SC) | Attorney-closing state | | Tennessee (TN) | Indebtedness tax $1.15/$1,000 | | Texas (TX) | No mortgage recording tax | | Utah (UT) | No mortgage tax |
This skill is part of the mortgage plugin. Install via:
/plugin marketplace add lendtrain/mortgage
/plugin install mortgage@mortgageFull source: github.com/lendtrain/mortgage
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 9,216 | 4,970 | -46% | 1 | 1 | 0% | 1,691 | 1,408 | -17% | 0 | 0 | — |
case-02 | pass→pass | 6,175 | 3,391 | -45% | 1 | 1 | 0% | 1,303 | 1,152 | -12% | 0 | 0 | — |
case-03 | fail→pass | 6,394 | 3,301 | -48% | 1 | 1 | 0% | 1,159 | 1,097 | -5% | 0 | 0 | — |
case-04 | pass→pass | 8,586 | 5,400 | -37% | 1 | 1 | 0% | 1,541 | 1,367 | -11% | 0 | 0 | — |
case-05 | pass→pass | 15,874 | 13,802 | -13% | 1 | 1 | 0% | 2,884 | 3,041 | +5% | 0 | 0 | — |
case-06 | pass→pass | 9,531 | 7,837 | -18% | 1 | 1 | 0% | 1,926 | 1,975 | +3% | 0 | 0 | — |
case-07 | pass→pass | 8,451 | 6,289 | -26% | 1 | 1 | 0% | 1,485 | 1,541 | +4% | 0 | 0 | — |
case-08 | pass→pass | 4,190 | 3,146 | -25% | 1 | 1 | 0% | 793 | 978 | +23% | 0 | 0 | — |
case-09 | pass→pass | 7,142 | 4,387 | -39% | 1 | 1 | 0% | 1,313 | 1,206 | -8% | 0 | 0 | — |
case-10 | pass→pass | 18,717 | 11,477 | -39% | 1 | 1 | 0% | 3,006 | 2,657 | -12% | 0 | 0 | — |
case-11 | fail→pass | 10,759 | 9,896 | -8% | 1 | 1 | 0% | 2,025 | 2,319 | +15% | 0 | 0 | — |
case-12 | pass→pass | 2,908 | 3,850 | +32% | 1 | 1 | 0% | 462 | 1,040 | +125% | 0 | 0 | — |
case-13 | pass→pass | 10,538 | 6,003 | -43% | 1 | 1 | 0% | 2,054 | 1,737 | -15% | 0 | 0 | — |
case-14 | fail→pass | 7,668 | 6,933 | -10% | 1 | 1 | 0% | 1,464 | 1,695 | +16% | 0 | 0 | — |
case-15 | pass→pass | 9,246 | 8,465 | -8% | 1 | 1 | 0% | 1,721 | 1,994 | +16% | 0 | 0 | — |
case-16 | pass→pass | 8,425 | 6,172 | -27% | 1 | 1 | 0% | 1,539 | 1,553 | +1% | 0 | 0 | — |
case-17 | pass→pass | 6,500 | 4,623 | -29% | 1 | 1 | 0% | 1,146 | 1,308 | +14% | 0 | 0 | — |
case-18 | pass→fail | 6,821 | 4,341 | -36% | 1 | 1 | 0% | 1,235 | 1,265 | +2% | 0 | 0 | — |
case-19 | fail→pass | 7,707 | 5,605 | -27% | 1 | 1 | 0% | 1,594 | 1,565 | -2% | 0 | 0 | — |
case-20 | fail→fail | 9,976 | 5,871 | -41% | 1 | 1 | 0% | 1,886 | 1,612 | -15% | 0 | 0 | — |
case-21 | pass→pass | 6,527 | 7,813 | +20% | 1 | 1 | 0% | 1,062 | 1,777 | +67% | 0 | 0 | — |
case-22 | pass→pass | 4,456 | 5,605 | +26% | 1 | 1 | 0% | 831 | 1,493 | +80% | 0 | 0 | — |
case-23 | pass→pass | 6,991 | 6,768 | -3% | 1 | 1 | 0% | 1,625 | 1,921 | +18% | 0 | 0 | — |
case-24 | pass→pass | 12,769 | 9,063 | -29% | 1 | 1 | 0% | 2,484 | 2,150 | -13% | 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. 24 cases were attempted. The headline lift of +13 percentage points is the difference between those two pass rates over the 24 comparable cases. 1 case got worse with the skill loaded, and it is 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.