Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Aggregate and display system metrics with anomaly detection for a time period
.claude/skills/ruvnet-observe-metrics/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -80% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -17% | 0% |
Aggregate counters, gauges, and histograms from the observability namespace and flag anomalies.
When you need a snapshot of system health -- task completion rates, error rates, active agent counts, memory usage, and token consumption. Useful for monitoring swarm performance and detecting degradation.
mcp__plugin_ruflo-core_ruflo__memory_search --namespace observability (or memory_list) to fetch metric records for the specified period (default: 1 hour). The memory_* tool family routes by namespace; agentdb_hierarchical-* does NOT, so use memory_* here.mcp__plugin_ruflo-core_ruflo__agentdb_pattern-search (ReasoningBank-routed; don't pass a namespace argument — pattern- tools ignore it) to establish baseline values for each metric.mcp__plugin_ruflo-core_ruflo__agentdb_pattern-store with type: 'metric-snapshot'. No namespace arg.mcp__plugin_ruflo-core_ruflo__memory_store --namespace observability for the snapshot tied to a timestamp.bashnpx @claude-flow/cli@latest memory search --query "system metrics for last hour" --namespace observability
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,239 | 4,626 | -50% | 1 | 1 | 0% | 1,990 | 859 | -57% | 0 | 0 | — |
case-02 | fail→fail | 12,307 | 7,820 | -36% | 1 | 1 | 0% | 2,324 | 928 | -60% | 0 | 0 | — |
case-03 | fail→fail | 15,552 | 4,444 | -71% | 1 | 1 | 0% | 4,039 | 890 | -78% | 0 | 0 | — |
case-04 | fail→pass | 8,422 | 2,850 | -66% | 1 | 1 | 0% | 1,606 | 1,127 | -30% | 0 | 0 | — |
case-05 | fail→pass | 5,167 | 1,441 | -72% | 1 | 1 | 0% | 993 | 781 | -21% | 0 | 0 | — |
case-06 | pass→pass | 10,403 | 1,197 | -88% | 1 | 1 | 0% | 2,077 | 673 | -68% | 0 | 0 | — |
case-07 | fail→pass | 2,971 | 1,924 | -35% | 1 | 1 | 0% | 596 | 913 | +53% | 0 | 0 | — |
case-08 | fail→fail | 10,356 | 1,542 | -85% | 1 | 1 | 0% | 2,071 | 762 | -63% | 0 | 0 | — |
case-09 | pass→pass | 12,674 | 5,696 | -55% | 1 | 1 | 0% | 2,432 | 1,510 | -38% | 0 | 0 | — |
case-10 | pass→pass | 12,443 | 7,183 | -42% | 1 | 1 | 0% | 2,318 | 1,988 | -14% | 0 | 0 | — |
case-11 | fail→fail | 6,722 | 1,614 | -76% | 1 | 1 | 0% | 1,188 | 782 | -34% | 0 | 0 | — |
case-12 | pass→pass | 7,592 | 2,241 | -70% | 1 | 1 | 0% | 1,435 | 908 | -37% | 0 | 0 | — |
case-13 | fail→pass | 19,789 | 1,278 | -94% | 1 | 1 | 0% | 3,640 | 719 | -80% | 0 | 0 | — |
case-14 | fail→pass | 4,848 | 1,492 | -69% | 1 | 1 | 0% | 861 | 715 | -17% | 0 | 0 | — |
case-15 | pass→pass | 8,119 | 1,027 | -87% | 1 | 1 | 0% | 1,552 | 664 | -57% | 0 | 0 | — |
case-16 | fail→pass | 6,798 | 1,987 | -71% | 1 | 1 | 0% | 1,222 | 867 | -29% | 0 | 0 | — |
case-17 | fail→pass | 11,728 | 1,769 | -85% | 1 | 1 | 0% | 2,213 | 785 | -65% | 0 | 0 | — |
case-18 | fail→pass | 9,587 | 1,864 | -81% | 1 | 1 | 0% | 1,951 | 829 | -58% | 0 | 0 | — |
case-19 | fail→pass | 11,113 | 1,801 | -84% | 1 | 1 | 0% | 2,327 | 822 | -65% | 0 | 0 | — |
case-20 | fail→fail | 7,846 | 3,643 | -54% | 1 | 1 | 0% | 1,485 | 1,167 | -21% | 0 | 0 | — |
case-21 | pass→pass | 4,205 | 2,021 | -52% | 1 | 1 | 0% | 731 | 786 | +8% | 0 | 0 | — |
case-22 | pass→pass | 11,905 | 7,138 | -40% | 1 | 1 | 0% | 2,484 | 1,749 | -30% | 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, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 19 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.