Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build and deploy the same feature consistently across web, mobile, and desktop platforms using API-first architecture and parallel implementation strategies.
.claude/skills/dokhacgiakhoa-multi-platform-apps-multi-platform/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -9% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 31% | 0% |
Build and deploy the same feature consistently across web, mobile, and desktop platforms using API-first architecture and parallel implementation strategies.
Extended thinking: This workflow orchestrates multiple specialized agents to ensure feature parity across platforms while maintaining platform-specific optimizations. The coordination strategy emphasizes shared contracts and parallel development with regular synchronization points. By establishing API contracts and data models upfront, teams can work independently while ensuring consistency. The workflow benefits include faster time-to-market, reduced integration issues, and maintainable cross-platform codebases.]
resources/implementation-playbook.md.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 39,673 | 34,384 | -13% | 1 | 1 | 0% | 8,254 | 4,230 | -49% | 0 | 0 | — |
case-02 | fail→fail | 32,220 | 26,311 | -18% | 1 | 1 | 0% | 5,235 | 5,473 | +5% | 0 | 0 | — |
case-03 | fail→fail | 42,462 | 40,779 | -4% | 1 | 1 | 0% | 8,244 | 8,426 | +2% | 0 | 0 | — |
case-04 | pass→pass | 18,607 | 19,577 | +5% | 1 | 1 | 0% | 2,900 | 3,809 | +31% | 0 | 0 | — |
case-05 | pass→pass | 23,380 | 22,074 | -6% | 1 | 1 | 0% | 3,831 | 4,564 | +19% | 0 | 0 | — |
case-06 | pass→pass | 18,954 | 21,121 | +11% | 1 | 1 | 0% | 3,240 | 4,118 | +27% | 0 | 0 | — |
case-07 | fail→fail | 14,777 | 10,925 | -26% | 1 | 1 | 0% | 2,279 | 2,350 | +3% | 0 | 0 | — |
case-08 | fail→pass | 19,639 | 23,232 | +18% | 1 | 1 | 0% | 3,339 | 4,543 | +36% | 0 | 0 | — |
case-09 | fail→fail | 17,209 | 19,434 | +13% | 1 | 1 | 0% | 2,693 | 3,794 | +41% | 0 | 0 | — |
case-10 | fail→pass | 16,572 | 11,610 | -30% | 1 | 1 | 0% | 2,724 | 2,467 | -9% | 0 | 0 | — |
case-11 | fail→fail | 22,725 | 17,827 | -22% | 1 | 1 | 0% | 4,422 | 4,064 | -8% | 0 | 0 | — |
case-12 | fail→fail | 15,594 | 7,134 | -54% | 1 | 1 | 0% | 2,602 | 1,765 | -32% | 0 | 0 | — |
case-13 | pass→pass | 12,730 | 12,763 | +0% | 1 | 1 | 0% | 1,999 | 2,934 | +47% | 0 | 0 | — |
case-14 | fail→fail | 19,033 | 17,181 | -10% | 1 | 1 | 0% | 3,332 | 3,999 | +20% | 0 | 0 | — |
case-15 | fail→pass | 17,168 | 16,367 | -5% | 1 | 1 | 0% | 2,745 | 3,372 | +23% | 0 | 0 | — |
case-16 | fail→fail | 17,695 | 17,404 | -2% | 1 | 1 | 0% | 2,666 | 3,645 | +37% | 0 | 0 | — |
case-17 | fail→fail | 16,831 | 16,590 | -1% | 1 | 1 | 0% | 2,572 | 3,525 | +37% | 0 | 0 | — |
case-18 | pass→pass | 16,950 | 16,817 | -1% | 1 | 1 | 0% | 2,812 | 3,753 | +33% | 0 | 0 | — |
case-19 | fail→pass | 18,237 | 21,493 | +18% | 1 | 1 | 0% | 3,158 | 4,947 | +57% | 0 | 0 | — |
case-20 | fail→fail | 20,356 | 18,537 | -9% | 1 | 1 | 0% | 2,872 | 3,687 | +28% | 0 | 0 | — |
case-21 | fail→fail | 24,100 | 18,318 | -24% | 1 | 1 | 0% | 3,518 | 4,041 | +15% | 0 | 0 | — |
case-22 | fail→fail | 12,765 | 16,387 | +28% | 1 | 1 | 0% | 2,053 | 3,666 | +79% | 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 +18 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.