Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Comprehensive guide for Firebase Crashlytics, including provisioning and SDK usage. Use this skill when the user needs help setting up Crashlytics, adding crash reporting, or using the Crashlytics SDK in their application.
.claude/skills/firebase-crashlytics/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -56% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -68% | 0% |
This skill provides a complete guide for getting started with Crashlytics on Android or iOS. Crash data collected from client applications can be read using the MCP server in the Firebase CLI.
Provisioning Crashlytics requires both a Firebase project and a Firebase app, either Android or iOS. To read the data collected by Crashlytics, install the MCP server in the Firebase CLI. See the firebase-basics skill for references.
To learn how to setup Crashlytics in your application code, choose your platform:
The SDK provides a number of features to make crash reports more actionable.
To learn how to customize crash reports and add additional debugging data, consult the documentation for your platform.
Customize Crash Reports for Android
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,044 | 3,992 | -56% | 1 | 1 | 0% | 1,637 | 1,128 | -31% | 0 | 0 | — |
case-02 | fail→pass | 5,268 | 3,861 | -27% | 1 | 1 | 0% | 909 | 1,016 | +12% | 0 | 0 | — |
case-03 | fail→pass | 11,082 | 3,994 | -64% | 1 | 1 | 0% | 2,026 | 1,088 | -46% | 0 | 0 | — |
case-04 | fail→pass | 7,923 | 2,034 | -74% | 1 | 1 | 0% | 1,361 | 595 | -56% | 0 | 0 | — |
case-05 | fail→pass | 11,342 | 2,582 | -77% | 1 | 1 | 0% | 2,129 | 680 | -68% | 0 | 0 | — |
case-06 | fail→pass | 12,351 | 1,685 | -86% | 1 | 1 | 0% | 2,126 | 574 | -73% | 0 | 0 | — |
case-07 | fail→pass | 4,968 | 2,552 | -49% | 1 | 1 | 0% | 704 | 651 | -8% | 0 | 0 | — |
case-08 | fail→pass | 5,132 | 1,691 | -67% | 1 | 1 | 0% | 895 | 569 | -36% | 0 | 0 | — |
case-09 | fail→pass | 6,212 | 2,301 | -63% | 1 | 1 | 0% | 1,228 | 694 | -43% | 0 | 0 | — |
case-10 | pass→pass | 9,593 | 2,849 | -70% | 1 | 1 | 0% | 1,565 | 756 | -52% | 0 | 0 | — |
case-11 | fail→pass | 8,679 | 1,517 | -83% | 1 | 1 | 0% | 1,541 | 536 | -65% | 0 | 0 | — |
case-12 | fail→pass | 7,591 | 3,014 | -60% | 1 | 1 | 0% | 1,473 | 720 | -51% | 0 | 0 | — |
case-13 | pass→pass | 9,589 | 5,092 | -47% | 1 | 1 | 0% | 1,801 | 1,226 | -32% | 0 | 0 | — |
case-14 | pass→pass | 6,151 | 3,805 | -38% | 1 | 1 | 0% | 1,221 | 976 | -20% | 0 | 0 | — |
case-15 | fail→pass | 10,293 | 1,563 | -85% | 1 | 1 | 0% | 1,741 | 537 | -69% | 0 | 0 | — |
case-16 | fail→pass | 2,807 | 2,509 | -11% | 1 | 1 | 0% | 414 | 661 | +60% | 0 | 0 | — |
case-17 | fail→pass | 2,525 | 1,768 | -30% | 1 | 1 | 0% | 463 | 583 | +26% | 0 | 0 | — |
case-18 | pass→pass | 9,848 | 3,655 | -63% | 1 | 1 | 0% | 1,719 | 879 | -49% | 0 | 0 | — |
case-19 | pass→pass | 8,788 | 3,029 | -66% | 1 | 1 | 0% | 1,546 | 792 | -49% | 0 | 0 | — |
case-20 | fail→fail | 2,961 | 5,227 | +77% | 1 | 1 | 0% | 466 | 852 | +83% | 0 | 0 | — |
case-21 | fail→pass | 4,557 | 1,547 | -66% | 1 | 1 | 0% | 452 | 493 | +9% | 0 | 0 | — |
case-22 | pass→pass | 13,300 | 12,328 | -7% | 1 | 1 | 0% | 2,563 | 2,745 | +7% | 0 | 0 | — |
case-23 | pass→pass | 10,207 | 7,615 | -25% | 1 | 1 | 0% | 2,082 | 1,752 | -16% | 0 | 0 | — |
case-24 | pass→pass | 8,935 | 6,078 | -32% | 1 | 1 | 0% | 1,611 | 1,427 | -11% | 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. 24 cases were attempted. The headline lift of +63 percentage points is the difference between those two pass rates over the 24 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/24/2026 | +50% |
Other measured skills in the registry, with their headline benchmark lift.