Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Convert SOAP web services to REST APIs with WSDL parsing and resource modeling
.claude/skills/a5c-ai-soap-to-rest-converter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -25% | 0% |
Converts SOAP web services to REST APIs, handling WSDL parsing, operation mapping, and RESTful resource modeling.
Enable SOAP modernization for:
| Tool | Purpose | Integration Method | |------|---------|-------------------| | SOAP UI | WSDL analysis | GUI | | wsdl2rest | Conversion | CLI | | Custom transformers | Type mapping | Library |
json{ "conversionId": "string", "timestamp": "ISO8601", "source": { "wsdl": "string", "operations": "number" }, "target": { "openapi": "string", "endpoints": "number" }, "mappings": [ { "soapOperation": "string", "restEndpoint": "string", "method": "string", "notes": "string" } ], "typeConversions": [], "manualReview": [] }
openapi-generator: REST spec generationapi-inventory-scanner: Endpoint discoveryapi-modernization-architect: API design| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,472 | 10,935 | +15% | 1 | 1 | 0% | 2,353 | 2,712 | +15% | 0 | 0 | — |
case-02 | fail→pass | 16,054 | 12,470 | -22% | 1 | 1 | 0% | 2,783 | 2,654 | -5% | 0 | 0 | — |
case-03 | fail→pass | 25,379 | 13,376 | -47% | 1 | 1 | 0% | 5,499 | 3,300 | -40% | 0 | 0 | — |
case-04 | fail→fail | 7,081 | 14,218 | +101% | 1 | 1 | 0% | 1,675 | 3,263 | +95% | 0 | 0 | — |
case-05 | fail→fail | 20,352 | 16,986 | -17% | 1 | 1 | 0% | 4,204 | 4,422 | +5% | 0 | 0 | — |
case-06 | fail→fail | 9,702 | 12,116 | +25% | 1 | 1 | 0% | 1,659 | 3,017 | +82% | 0 | 0 | — |
case-07 | fail→pass | 14,791 | 10,642 | -28% | 1 | 1 | 0% | 2,523 | 2,677 | +6% | 0 | 0 | — |
case-08 | fail→fail | 6,424 | 1,531 | -76% | 1 | 1 | 0% | 992 | 722 | -27% | 0 | 0 | — |
case-09 | fail→fail | 13,929 | 11,607 | -17% | 1 | 1 | 0% | 2,543 | 2,800 | +10% | 0 | 0 | — |
case-10 | fail→fail | 10,811 | 10,332 | -4% | 1 | 1 | 0% | 2,063 | 2,235 | +8% | 0 | 0 | — |
case-11 | fail→fail | 15,303 | 17,108 | +12% | 1 | 1 | 0% | 2,717 | 3,208 | +18% | 0 | 0 | — |
case-12 | fail→fail | 11,782 | 10,466 | -11% | 1 | 1 | 0% | 1,963 | 2,086 | +6% | 0 | 0 | — |
case-13 | fail→fail | 6,948 | 5,088 | -27% | 1 | 1 | 0% | 1,538 | 1,529 | -1% | 0 | 0 | — |
case-14 | pass→pass | 5,612 | 5,589 | -0% | 1 | 1 | 0% | 1,161 | 1,592 | +37% | 0 | 0 | — |
case-15 | pass→pass | 9,439 | 7,113 | -25% | 1 | 1 | 0% | 2,071 | 2,041 | -1% | 0 | 0 | — |
case-16 | fail→fail | 4,730 | 3,873 | -18% | 1 | 1 | 0% | 987 | 1,231 | +25% | 0 | 0 | — |
case-17 | fail→fail | 12,373 | 12,563 | +2% | 1 | 1 | 0% | 2,275 | 2,932 | +29% | 0 | 0 | — |
case-18 | fail→fail | 10,296 | 2,720 | -74% | 1 | 1 | 0% | 1,690 | 835 | -51% | 0 | 0 | — |
case-19 | fail→pass | 6,285 | 1,620 | -74% | 1 | 1 | 0% | 1,005 | 755 | -25% | 0 | 0 | — |
case-20 | fail→pass | 6,351 | 1,717 | -73% | 1 | 1 | 0% | 1,037 | 729 | -30% | 0 | 0 | — |
case-21 | fail→fail | 12,940 | 11,941 | -8% | 1 | 1 | 0% | 2,322 | 2,854 | +23% | 0 | 0 | — |
case-22 | fail→fail | 7,579 | 6,159 | -19% | 1 | 1 | 0% | 1,520 | 1,697 | +12% | 0 | 0 | — |
case-23 | fail→pass | 10,243 | 2,859 | -72% | 1 | 1 | 0% | 1,673 | 1,041 | -38% | 0 | 0 | — |
case-24 | fail→pass | 11,906 | 2,131 | -82% | 1 | 1 | 0% | 1,997 | 920 | -54% | 0 | 0 | — |
case-25 | fail→fail | 5,354 | 2,902 | -46% | 1 | 1 | 0% | 862 | 984 | +14% | 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. The headline lift of +32 percentage points is the difference between those two pass rates over the 25 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.