Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Integrate Retrofit, OkHttp, and Kotlinx Serialization for type-safe API communication in Android. Use when building API clients, adding interceptors, or configuring network security—not standalone MockWebServer or API-test setup.
.claude/skills/hoangnguyen0403-android-networking/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -41% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 3% | 0% |
@SerialName for JSON field mapping.See setup & wrappers for DTO and API examples.
suspend functions.See setup & wrappers for API endpoint definitions.
Bearer token injection and HttpLoggingInterceptor (debug only).Result wrapper or Either in Repository layer.isMinifyEnabled = true.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 11,082 | 8,325 | -25% | 1 | 1 | 0% | 2,353 | 2,141 | -9% | 0 | 0 | — |
case-02 | pass→pass | 19,898 | 7,721 | -61% | 1 | 1 | 0% | 3,481 | 2,057 | -41% | 0 | 0 | — |
case-03 | pass→pass | 14,680 | 12,720 | -13% | 1 | 1 | 0% | 3,052 | 3,147 | +3% | 0 | 0 | — |
case-04 | pass→pass | 7,677 | 4,519 | -41% | 1 | 1 | 0% | 1,558 | 1,169 | -25% | 0 | 0 | — |
case-05 | pass→pass | 7,660 | 5,325 | -30% | 1 | 1 | 0% | 1,512 | 1,217 | -20% | 0 | 0 | — |
case-06 | fail→pass | 12,676 | 5,970 | -53% | 1 | 1 | 0% | 2,304 | 1,582 | -31% | 0 | 0 | — |
case-20 | pass→pass | 12,095 | 8,983 | -26% | 1 | 1 | 0% | 2,465 | 2,206 | -11% | 0 | 0 | — |
case-07 | pass→pass | 17,044 | 7,152 | -58% | 1 | 1 | 0% | 2,593 | 1,619 | -38% | 0 | 0 | — |
case-08 | fail→pass | 15,118 | 9,671 | -36% | 1 | 1 | 0% | 2,947 | 2,233 | -24% | 0 | 0 | — |
case-09 | pass→pass | 17,296 | 6,802 | -61% | 1 | 1 | 0% | 2,943 | 1,684 | -43% | 0 | 0 | — |
case-10 | pass→pass | 11,950 | 7,780 | -35% | 1 | 1 | 0% | 2,334 | 1,833 | -21% | 0 | 0 | — |
case-21 | pass→pass | 16,447 | 8,672 | -47% | 1 | 1 | 0% | 2,524 | 1,917 | -24% | 0 | 0 | — |
case-11 | pass→pass | 9,123 | 6,438 | -29% | 1 | 1 | 0% | 2,022 | 1,604 | -21% | 0 | 0 | — |
case-12 | pass→pass | 13,159 | 7,306 | -44% | 1 | 1 | 0% | 2,571 | 1,745 | -32% | 0 | 0 | — |
case-13 | fail→pass | 16,003 | 12,776 | -20% | 1 | 1 | 0% | 3,093 | 3,250 | +5% | 0 | 0 | — |
case-14 | pass→pass | 10,736 | 6,814 | -37% | 1 | 1 | 0% | 1,861 | 1,402 | -25% | 0 | 0 | — |
case-15 | pass→pass | 10,353 | 4,368 | -58% | 1 | 1 | 0% | 1,993 | 1,185 | -41% | 0 | 0 | — |
case-16 | pass→pass | 11,507 | 9,388 | -18% | 1 | 1 | 0% | 2,421 | 2,345 | -3% | 0 | 0 | — |
case-17 | pass→pass | 15,122 | 10,842 | -28% | 1 | 1 | 0% | 3,056 | 2,196 | -28% | 0 | 0 | — |
case-18 | pass→pass | 6,003 | 2,532 | -58% | 1 | 1 | 0% | 1,139 | 807 | -29% | 0 | 0 | — |
case-19 | pass→pass | 10,330 | 5,660 | -45% | 1 | 1 | 0% | 1,919 | 1,440 | -25% | 0 | 0 | — |
case-22 | pass→pass | 18,472 | 14,213 | -23% | 1 | 1 | 0% | 3,197 | 2,824 | -12% | 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. The headline lift of +14 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.