Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Manage service account manager operations. Auto-activating skill for GCP Skills. Triggers on: service account manager, service account manager Part of the GCP Skills skill category. Use when working with service account manager functionality. Trigger with phrases like "service account manager", "service manager", "service".
.claude/skills/dicklesworthstone-service-account-manager/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-22 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 66% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 15% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 25% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 21% | 0% |
This skill provides automated assistance for service account manager tasks within the GCP Skills domain.
This skill activates automatically when you:
Example: Basic Usage Request: "Help me with service account manager" Result: Provides step-by-step guidance and generates appropriate configurations
| Error | Cause | Solution | |-------|-------|----------| | Configuration invalid | Missing required fields | Check documentation for required parameters | | Tool not found | Dependency not installed | Install required tools per prerequisites | | Permission denied | Insufficient access | Verify credentials and permissions |
Part of the GCP Skills skill category. Tags: gcp, bigquery, vertex-ai, cloud-run, firebase
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 11,546 | 13,215 | +14% | 1 | 1 | 0% | 1,092 | 1,817 | +66% | 0 | 0 | — |
case-01 | pass→pass | 17,606 | 17,015 | -3% | 1 | 1 | 0% | 2,340 | 2,681 | +15% | 0 | 0 | — |
case-02 | pass→pass | 16,267 | 17,159 | +5% | 1 | 1 | 0% | 2,209 | 2,768 | +25% | 0 | 0 | — |
case-03 | pass→pass | 18,232 | 19,597 | +7% | 1 | 1 | 0% | 2,531 | 3,066 | +21% | 0 | 0 | — |
case-05 | pass→pass | 11,565 | 11,695 | +1% | 1 | 1 | 0% | 1,156 | 1,520 | +31% | 0 | 0 | — |
case-06 | pass→pass | 10,311 | 10,474 | +2% | 1 | 1 | 0% | 916 | 1,421 | +55% | 0 | 0 | — |
case-07 | pass→pass | 14,497 | 17,490 | +21% | 1 | 1 | 0% | 1,654 | 2,515 | +52% | 0 | 0 | — |
case-08 | pass→pass | 20,352 | 21,996 | +8% | 1 | 1 | 0% | 2,475 | 3,025 | +22% | 0 | 0 | — |
case-09 | pass→pass | 15,103 | 14,216 | -6% | 1 | 1 | 0% | 1,613 | 2,018 | +25% | 0 | 0 | — |
case-10 | pass→pass | 13,070 | 14,461 | +11% | 1 | 1 | 0% | 1,524 | 2,017 | +32% | 0 | 0 | — |
case-11 | pass→pass | 17,922 | 19,685 | +10% | 1 | 1 | 0% | 2,131 | 3,057 | +43% | 0 | 0 | — |
case-12 | pass→pass | 11,955 | 11,905 | -0% | 1 | 1 | 0% | 1,087 | 1,581 | +45% | 0 | 0 | — |
case-13 | pass→pass | 19,163 | 18,470 | -4% | 1 | 1 | 0% | 2,189 | 2,608 | +19% | 0 | 0 | — |
case-14 | pass→pass | 15,822 | 16,798 | +6% | 1 | 1 | 0% | 2,076 | 2,605 | +25% | 0 | 0 | — |
case-15 | pass→pass | 17,349 | 19,796 | +14% | 1 | 1 | 0% | 2,264 | 3,198 | +41% | 0 | 0 | — |
case-16 | pass→pass | 15,272 | 16,386 | +7% | 1 | 1 | 0% | 1,729 | 2,284 | +32% | 0 | 0 | — |
case-17 | pass→pass | 22,164 | 24,243 | +9% | 1 | 1 | 0% | 2,741 | 3,831 | +40% | 0 | 0 | — |
case-18 | pass→pass | 19,140 | 18,426 | -4% | 1 | 1 | 0% | 2,269 | 2,586 | +14% | 0 | 0 | — |
case-19 | pass→pass | 19,847 | 19,702 | -1% | 1 | 1 | 0% | 2,805 | 3,357 | +20% | 0 | 0 | — |
case-20 | pass→pass | 18,132 | 17,189 | -5% | 1 | 1 | 0% | 2,467 | 2,641 | +7% | 0 | 0 | — |
case-21 | fail→fail | 11,661 | 10,067 | -14% | 1 | 1 | 0% | 1,295 | 1,269 | -2% | 0 | 0 | — |
case-22 | fail→pass | 18,794 | 18,136 | -4% | 1 | 1 | 0% | 2,423 | 2,658 | +10% | 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 +5 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.