Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Install and migrate to Jetpack Navigation 3. Use when implementing Navigation 3 patterns including NavDisplay, NavKey routes, deep links, multiple backstacks, scenes (dialogs, bottom sheets), or migrating from Navigation 2.
.claude/skills/hoangnguyen0403-android-navigation-3/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -26% | 0% |
Guide for implementing and migrating to Navigation 3 in Jetpack Compose.
Navigation 3 replaces the previous NavHost/NavController pattern with a simpler, state-driven approach:
mutableStateListOf<Any>.NavDisplay renders the current route based on a lambda.kotlinval backStack = remember { mutableStateListOf<Any>(RouteHome) } NavDisplay( backStack = backStack, onBack = { backStack.removeLastOrNull() }, entryProvider = { key -> when (key) { is RouteHome -> NavEntry(key) { HomeScreen(onNavigate = { backStack.add(it) }) } is RouteDetail -> NavEntry(key) { DetailScreen(key.id) } else -> error("Unknown route: $key") } } )
See migration guide for step-by-step conversion from NavHost/NavController to NavDisplay.
Key changes:
NavHost with NavDisplay.NavController.navigate() with direct list manipulation.navArgument with data class properties.See recipes for code examples:
mutableStateListOf<Any>.NavDisplay handles all routes in entryProvider../gradlew build succeeds.NavDisplay with a state list.remember { navController() }: Navigation 3 doesn't use NavController.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→pass | 17,592 | 7,435 | -58% | 1 | 1 | 0% | 2,861 | 2,134 | -25% | 0 | 0 | — |
case-05 | pass→pass | 13,193 | 10,270 | -22% | 1 | 1 | 0% | 2,456 | 2,224 | -9% | 0 | 0 | — |
case-01 | fail→pass | 14,226 | 6,741 | -53% | 1 | 1 | 0% | 3,005 | 2,143 | -29% | 0 | 0 | — |
case-02 | fail→pass | 14,667 | 8,126 | -45% | 1 | 1 | 0% | 3,077 | 2,302 | -25% | 0 | 0 | — |
case-03 | fail→pass | 10,905 | 8,532 | -22% | 1 | 1 | 0% | 2,381 | 2,518 | +6% | 0 | 0 | — |
case-06 | fail→pass | 14,676 | 6,883 | -53% | 1 | 1 | 0% | 2,547 | 1,887 | -26% | 0 | 0 | — |
case-07 | fail→pass | 9,802 | 5,325 | -46% | 1 | 1 | 0% | 1,803 | 1,640 | -9% | 0 | 0 | — |
case-08 | fail→pass | 11,700 | 10,233 | -13% | 1 | 1 | 0% | 2,389 | 2,709 | +13% | 0 | 0 | — |
case-09 | fail→pass | 15,490 | 8,724 | -44% | 1 | 1 | 0% | 3,068 | 2,240 | -27% | 0 | 0 | — |
case-10 | fail→pass | 16,637 | 13,478 | -19% | 1 | 1 | 0% | 2,926 | 2,884 | -1% | 0 | 0 | — |
case-11 | fail→pass | 16,722 | 9,689 | -42% | 1 | 1 | 0% | 2,878 | 2,730 | -5% | 0 | 0 | — |
case-12 | fail→pass | 18,296 | 14,748 | -19% | 1 | 1 | 0% | 3,153 | 3,204 | +2% | 0 | 0 | — |
case-13 | pass→pass | 13,861 | 11,404 | -18% | 1 | 1 | 0% | 2,626 | 2,906 | +11% | 0 | 0 | — |
case-14 | fail→fail | 12,793 | 10,449 | -18% | 1 | 1 | 0% | 2,322 | 2,642 | +14% | 0 | 0 | — |
case-15 | fail→pass | 9,968 | 6,273 | -37% | 1 | 1 | 0% | 2,005 | 1,816 | -9% | 0 | 0 | — |
case-16 | fail→pass | 10,218 | 4,470 | -56% | 1 | 1 | 0% | 1,712 | 1,281 | -25% | 0 | 0 | — |
case-17 | fail→pass | 13,973 | 9,850 | -30% | 1 | 1 | 0% | 2,442 | 2,468 | +1% | 0 | 0 | — |
case-18 | pass→pass | 11,683 | 8,178 | -30% | 1 | 1 | 0% | 2,163 | 2,154 | -0% | 0 | 0 | — |
case-19 | pass→pass | 10,017 | 8,649 | -14% | 1 | 1 | 0% | 2,078 | 2,363 | +14% | 0 | 0 | — |
case-20 | pass→pass | 9,977 | 7,536 | -24% | 1 | 1 | 0% | 2,152 | 2,224 | +3% | 0 | 0 | — |
case-21 | pass→pass | 10,161 | 7,509 | -26% | 1 | 1 | 0% | 2,211 | 2,204 | -0% | 0 | 0 | — |
case-22 | fail→pass | 12,254 | 6,625 | -46% | 1 | 1 | 0% | 2,283 | 1,815 | -20% | 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 +68 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.