Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Integrate OpenTelemetry tracing and metrics into SDKs
.claude/skills/a5c-ai-opentelemetry-integrator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 95% | 0% |
This skill integrates OpenTelemetry observability into SDKs, providing distributed tracing, metrics collection, and context propagation for comprehensive API monitoring.
yamlskill: name: opentelemetry-integrator context: tracing: enabled: true propagator: w3c-trace-context sampling: parentBased sampleRate: 0.1 metrics: enabled: true exportInterval: 30s metrics: - sdk.request.duration - sdk.request.count - sdk.error.count exporters: traces: otlp metrics: prometheus serviceName: "my-sdk"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 20,623 | 23,288 | +13% | 1 | 1 | 0% | 3,225 | 4,517 | +40% | 0 | 0 | — |
case-02 | fail→fail | 10,532 | 10,150 | -4% | 1 | 1 | 0% | 2,151 | 2,627 | +22% | 0 | 0 | — |
case-03 | pass→pass | 14,004 | 14,291 | +2% | 1 | 1 | 0% | 1,814 | 3,536 | +95% | 0 | 0 | — |
case-04 | pass→pass | 16,984 | 23,931 | +41% | 1 | 1 | 0% | 3,557 | 4,509 | +27% | 0 | 0 | — |
case-05 | fail→fail | 19,930 | 21,577 | +8% | 1 | 1 | 0% | 2,537 | 3,340 | +32% | 0 | 0 | — |
case-06 | pass→pass | 22,079 | 22,093 | +0% | 1 | 1 | 0% | 3,161 | 3,685 | +17% | 0 | 0 | — |
case-07 | pass→pass | 15,516 | 14,530 | -6% | 1 | 1 | 0% | 1,936 | 2,221 | +15% | 0 | 0 | — |
case-08 | pass→pass | 44,772 | 12,701 | -72% | 1 | 1 | 0% | 3,941 | 1,809 | -54% | 0 | 0 | — |
case-09 | fail→fail | 18,088 | 21,212 | +17% | 1 | 1 | 0% | 3,848 | 3,968 | +3% | 0 | 0 | — |
case-10 | fail→fail | 13,246 | 9,980 | -25% | 1 | 1 | 0% | 1,471 | 2,417 | +64% | 0 | 0 | — |
case-11 | pass→pass | 13,966 | 9,507 | -32% | 1 | 1 | 0% | 1,587 | 1,276 | -20% | 0 | 0 | — |
case-12 | fail→pass | 13,257 | 10,868 | -18% | 1 | 1 | 0% | 2,400 | 1,515 | -37% | 0 | 0 | — |
case-13 | pass→pass | 18,378 | 17,860 | -3% | 1 | 1 | 0% | 2,781 | 2,925 | +5% | 0 | 0 | — |
case-14 | pass→pass | 17,117 | 14,446 | -16% | 1 | 1 | 0% | 2,488 | 3,488 | +40% | 0 | 0 | — |
case-15 | fail→pass | 14,159 | 7,782 | -45% | 1 | 1 | 0% | 1,659 | 1,862 | +12% | 0 | 0 | — |
case-16 | pass→pass | 11,074 | 5,753 | -48% | 1 | 1 | 0% | 1,907 | 1,462 | -23% | 0 | 0 | — |
case-17 | pass→pass | 22,149 | 22,142 | -0% | 1 | 1 | 0% | 3,140 | 3,481 | +11% | 0 | 0 | — |
case-18 | fail→pass | 20,359 | 7,284 | -64% | 1 | 1 | 0% | 2,641 | 805 | -70% | 0 | 0 | — |
case-19 | fail→pass | 14,085 | 7,940 | -44% | 1 | 1 | 0% | 1,558 | 862 | -45% | 0 | 0 | — |
case-20 | pass→pass | 7,675 | 12,787 | +67% | 1 | 1 | 0% | 1,531 | 1,974 | +29% | 0 | 0 | — |
case-21 | pass→pass | 12,554 | 15,772 | +26% | 1 | 1 | 0% | 2,827 | 2,704 | -4% | 0 | 0 | — |
case-22 | pass→pass | 10,397 | 7,992 | -23% | 1 | 1 | 0% | 2,050 | 1,958 | -4% | 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.