Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Implement health check and readiness endpoints for SDK consumers
.claude/skills/a5c-ai-health-check-endpoint/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 10% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -3% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -12% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 48% | 0% |
This skill implements health check and readiness endpoints that SDK consumers can use to verify connectivity and dependency health, especially important for Kubernetes and load balancer integration.
yamlskill: name: health-check-endpoint context: endpoints: health: /health ready: /ready live: /live checks: - name: database type: tcp critical: true - name: cache type: redis critical: false - name: externalApi type: http url: https://api.example.com/health kubernetes: livenessProbe: path: /live periodSeconds: 10 readinessProbe: path: /ready periodSeconds: 5
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 17,494 | 15,922 | -9% | 1 | 1 | 0% | 2,208 | 2,433 | +10% | 0 | 0 | — |
case-02 | fail→pass | 16,813 | 17,859 | +6% | 1 | 1 | 0% | 2,590 | 2,519 | -3% | 0 | 0 | — |
case-03 | pass→pass | 12,960 | 10,337 | -20% | 1 | 1 | 0% | 1,351 | 1,310 | -3% | 0 | 0 | — |
case-04 | pass→pass | 10,826 | 11,990 | +11% | 1 | 1 | 0% | 1,860 | 1,644 | -12% | 0 | 0 | — |
case-05 | pass→pass | 17,071 | 20,960 | +23% | 1 | 1 | 0% | 2,010 | 2,980 | +48% | 0 | 0 | — |
case-06 | pass→pass | 18,979 | 17,474 | -8% | 1 | 1 | 0% | 2,403 | 3,644 | +52% | 0 | 0 | — |
case-07 | pass→pass | 16,198 | 13,802 | -15% | 1 | 1 | 0% | 1,799 | 2,555 | +42% | 0 | 0 | — |
case-08 | pass→pass | 18,577 | 17,360 | -7% | 1 | 1 | 0% | 2,333 | 2,297 | -2% | 0 | 0 | — |
case-09 | pass→pass | 10,205 | 10,599 | +4% | 1 | 1 | 0% | 946 | 1,605 | +70% | 0 | 0 | — |
case-10 | pass→pass | 9,348 | 5,028 | -46% | 1 | 1 | 0% | 820 | 1,260 | +54% | 0 | 0 | — |
case-11 | pass→pass | 12,097 | 2,876 | -76% | 1 | 1 | 0% | 1,187 | 885 | -25% | 0 | 0 | — |
case-12 | pass→pass | 19,025 | 12,381 | -35% | 1 | 1 | 0% | 2,461 | 1,626 | -34% | 0 | 0 | — |
case-13 | pass→pass | 14,979 | 8,229 | -45% | 1 | 1 | 0% | 1,761 | 908 | -48% | 0 | 0 | — |
case-14 | pass→pass | 12,579 | 7,554 | -40% | 1 | 1 | 0% | 1,333 | 1,787 | +34% | 0 | 0 | — |
case-15 | pass→pass | 20,077 | 13,775 | -31% | 1 | 1 | 0% | 2,407 | 2,655 | +10% | 0 | 0 | — |
case-16 | pass→pass | 12,103 | 18,406 | +52% | 1 | 1 | 0% | 2,058 | 2,687 | +31% | 0 | 0 | — |
case-17 | pass→pass | 20,132 | 25,812 | +28% | 1 | 1 | 0% | 3,536 | 4,189 | +18% | 0 | 0 | — |
case-18 | pass→pass | 20,640 | 16,417 | -20% | 1 | 1 | 0% | 2,702 | 3,409 | +26% | 0 | 0 | — |
case-19 | pass→pass | 7,550 | 3,376 | -55% | 1 | 1 | 0% | 1,402 | 978 | -30% | 0 | 0 | — |
case-20 | pass→pass | 13,171 | 12,843 | -2% | 1 | 1 | 0% | 2,749 | 3,020 | +10% | 0 | 0 | — |
case-21 | pass→pass | 21,482 | 23,239 | +8% | 1 | 1 | 0% | 3,345 | 4,036 | +21% | 0 | 0 | — |
case-22 | pass→pass | 13,956 | 23,041 | +65% | 1 | 1 | 0% | 2,904 | 4,091 | +41% | 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.