Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyzes social media campaign performance across platforms with engagement metrics, ROI calculations, and audience insights for data-driven marketing decisions
.claude/skills/nicepkg-social-media-analyzer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-03 | ✓→✗ | ▼ Worse | -1% | 0% |
This skill provides comprehensive analysis of social media campaign performance, helping marketing agencies deliver actionable insights to clients.
Campaign data including:
Formats accepted:
Results include:
"Analyze this Facebook campaign data and calculate engagement metrics" "What's the ROI on this Instagram ad campaign with $500 spend and 2,000 clicks?" "Compare performance across all social platforms for the last month"
calculate_metrics.py: Core calculation engine for all social media metricsanalyze_performance.py: Performance analysis and recommendation generation| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 5,745 | 7,853 | +37% | 1 | 1 | 0% | 956 | 1,099 | +15% | 0 | 0 | — |
case-02 | pass→pass | 3,990 | 4,596 | +15% | 1 | 1 | 0% | 577 | 1,238 | +115% | 0 | 0 | — |
case-03 | pass→fail | 12,630 | 9,049 | -28% | 1 | 1 | 0% | 1,926 | 1,909 | -1% | 0 | 0 | — |
case-04 | pass→pass | 4,687 | 5,890 | +26% | 1 | 1 | 0% | 883 | 1,628 | +84% | 0 | 0 | — |
case-05 | pass→pass | 13,648 | 11,899 | -13% | 1 | 1 | 0% | 2,166 | 2,606 | +20% | 0 | 0 | — |
case-06 | pass→pass | 3,765 | 3,813 | +1% | 1 | 1 | 0% | 654 | 1,242 | +90% | 0 | 0 | — |
case-07 | pass→pass | 3,150 | 5,249 | +67% | 1 | 1 | 0% | 582 | 1,491 | +156% | 0 | 0 | — |
case-08 | pass→pass | 11,991 | 11,925 | -1% | 1 | 1 | 0% | 1,836 | 2,443 | +33% | 0 | 0 | — |
case-09 | pass→pass | 12,420 | 12,053 | -3% | 1 | 1 | 0% | 1,964 | 2,431 | +24% | 0 | 0 | — |
case-10 | pass→pass | 7,301 | 1,994 | -73% | 1 | 1 | 0% | 1,161 | 895 | -23% | 0 | 0 | — |
case-11 | fail→pass | 7,140 | 1,690 | -76% | 1 | 1 | 0% | 1,132 | 812 | -28% | 0 | 0 | — |
case-12 | pass→pass | 14,379 | 12,953 | -10% | 1 | 1 | 0% | 2,265 | 2,634 | +16% | 0 | 0 | — |
case-13 | pass→pass | 3,103 | 2,906 | -6% | 1 | 1 | 0% | 594 | 1,090 | +84% | 0 | 0 | — |
case-14 | pass→pass | 2,482 | 3,458 | +39% | 1 | 1 | 0% | 406 | 1,106 | +172% | 0 | 0 | — |
case-15 | fail→pass | 8,338 | 3,921 | -53% | 1 | 1 | 0% | 1,192 | 1,160 | -3% | 0 | 0 | — |
case-16 | pass→pass | 11,838 | 2,362 | -80% | 1 | 1 | 0% | 1,723 | 902 | -48% | 0 | 0 | — |
case-17 | pass→pass | 9,987 | 4,758 | -52% | 1 | 1 | 0% | 1,502 | 1,310 | -13% | 0 | 0 | — |
case-18 | pass→pass | 8,975 | 7,529 | -16% | 1 | 1 | 0% | 1,296 | 1,623 | +25% | 0 | 0 | — |
case-19 | fail→pass | 11,766 | 10,311 | -12% | 1 | 1 | 0% | 1,657 | 2,035 | +23% | 0 | 0 | — |
case-20 | pass→pass | 7,391 | 6,318 | -15% | 1 | 1 | 0% | 965 | 1,409 | +46% | 0 | 0 | — |
case-21 | pass→pass | 8,240 | 5,325 | -35% | 1 | 1 | 0% | 1,215 | 1,292 | +6% | 0 | 0 | — |
case-22 | pass→pass | 7,774 | 2,197 | -72% | 1 | 1 | 0% | 1,138 | 932 | -18% | 0 | 0 | — |
case-23 | pass→pass | 3,165 | 3,350 | +6% | 1 | 1 | 0% | 566 | 1,128 | +99% | 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. 23 cases were attempted. The headline lift of +13 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.