Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Comprehensive API testing for REST and GraphQL endpoints with contract validation
.claude/skills/a5c-ai-api-testing/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-11 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-04 | ✓→✓ | = Same ✓ | -7% | 0% |
| case-05 | ✓→✓ | = Same ✓ | -31% | 0% |
This skill provides expert-level capabilities for API testing, covering REST, GraphQL, authentication flows, schema validation, and contract testing integration.
api-testing.js - API test implementationcontract-testing.js - Contract testing workflowssecurity-testing.js - API security testingperformance-testing.js - API performance validationsupertest - HTTP assertions (Node.js)axios - HTTP clientajv - JSON schema validationgraphql-request - GraphQL clientjavascript{ kind: 'skill', skill: { name: 'api-testing', context: { action: 'execute-tests', baseUrl: 'https://api.example.com', specPath: 'openapi.yaml', authType: 'bearer', validateSchema: true } } }
The skill can auto-discover API specifications from OpenAPI/Swagger files and generate comprehensive test suites.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 11,746 | 20,047 | +71% | 1 | 1 | 0% | 2,095 | 3,757 | +79% | 0 | 0 | — |
case-02 | fail→fail | 31,965 | 24,223 | -24% | 1 | 1 | 0% | 3,334 | 3,455 | +4% | 0 | 0 | — |
case-03 | fail→fail | 15,668 | 22,049 | +41% | 1 | 1 | 0% | 2,110 | 2,443 | +16% | 0 | 0 | — |
case-04 | pass→pass | 15,197 | 12,347 | -19% | 1 | 1 | 0% | 1,976 | 1,834 | -7% | 0 | 0 | — |
case-05 | pass→pass | 13,333 | 8,632 | -35% | 1 | 1 | 0% | 1,486 | 1,023 | -31% | 0 | 0 | — |
case-06 | pass→pass | 12,296 | 6,617 | -46% | 1 | 1 | 0% | 1,259 | 652 | -48% | 0 | 0 | — |
case-07 | pass→pass | 6,659 | 1,840 | -72% | 1 | 1 | 0% | 1,202 | 736 | -39% | 0 | 0 | — |
case-08 | pass→pass | 17,018 | 16,654 | -2% | 1 | 1 | 0% | 1,884 | 2,691 | +43% | 0 | 0 | — |
case-09 | pass→pass | 10,049 | 6,688 | -33% | 1 | 1 | 0% | 1,960 | 730 | -63% | 0 | 0 | — |
case-10 | pass→pass | 16,253 | 8,791 | -46% | 1 | 1 | 0% | 2,004 | 2,142 | +7% | 0 | 0 | — |
case-11 | fail→pass | 7,238 | 7,952 | +10% | 1 | 1 | 0% | 1,289 | 1,030 | -20% | 0 | 0 | — |
case-12 | fail→pass | 10,837 | 11,666 | +8% | 1 | 1 | 0% | 1,163 | 1,004 | -14% | 0 | 0 | — |
case-13 | fail→pass | 11,023 | 8,494 | -23% | 1 | 1 | 0% | 1,010 | 1,027 | +2% | 0 | 0 | — |
case-14 | pass→pass | 14,692 | 10,930 | -26% | 1 | 1 | 0% | 1,707 | 1,525 | -11% | 0 | 0 | — |
case-15 | pass→pass | 16,732 | 1,713 | -90% | 1 | 1 | 0% | 2,020 | 671 | -67% | 0 | 0 | — |
case-16 | pass→pass | 13,073 | 10,289 | -21% | 1 | 1 | 0% | 2,214 | 1,205 | -46% | 0 | 0 | — |
case-17 | fail→fail | 10,016 | 3,033 | -70% | 1 | 1 | 0% | 992 | 978 | -1% | 0 | 0 | — |
case-18 | pass→pass | 9,054 | 8,926 | -1% | 1 | 1 | 0% | 1,639 | 1,042 | -36% | 0 | 0 | — |
case-19 | pass→pass | 18,248 | 11,112 | -39% | 1 | 1 | 0% | 2,228 | 1,478 | -34% | 0 | 0 | — |
case-20 | pass→pass | 12,798 | 13,552 | +6% | 1 | 1 | 0% | 1,455 | 1,732 | +19% | 0 | 0 | — |
case-21 | pass→pass | 10,045 | 9,592 | -5% | 1 | 1 | 0% | 907 | 1,315 | +45% | 0 | 0 | — |
case-22 | pass→pass | 10,471 | 9,146 | -13% | 1 | 1 | 0% | 1,096 | 1,292 | +18% | 0 | 0 | — |
case-23 | pass→pass | 16,444 | 4,619 | -72% | 1 | 1 | 0% | 1,849 | 1,274 | -31% | 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. 23 cases were attempted. The headline lift of +13 percentage points is the difference between those two pass rates over the 23 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.