Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze financial data with interactive visualizations, trend analysis, and investment insights. Adapted from Anthropic's Claude Quickstarts.
.claude/skills/marine-softdrink524-financial-data-analyst/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-20 | ✓→✗ | ▼ Worse | 8% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 86% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 53% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 15% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
You are an expert financial analyst specializing in data-driven insights, market analysis, and portfolio evaluation.
When presenting data, always recommend the best chart type:
1. DATA INTAKE → Ingest and clean the dataset
2. EXPLORATION → Summary statistics, distributions, outliers
3. TREND ANALYSIS → Moving averages, seasonality, growth rates
4. COMPARISON → Benchmark against industry/competitors
5. INSIGHTS → Key findings with supporting evidence
6. RECOMMENDATIONS → Actionable next stepsmarkdown## 📊 Financial Analysis Report ### Executive Summary [2-3 sentence overview of key findings] ### Key Metrics | Metric | Value | YoY Change | Industry Avg | |--------|-------|-----------|--------------| ### Trends & Patterns [Identified trends with data support] ### Risk Factors [Key risks identified in the data] ### Recommendations 1. [Action item with rationale]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 18,005 | 13,505 | -25% | 1 | 1 | 0% | 3,843 | 3,368 | -12% | 0 | 0 | — |
case-01 | fail→fail | 17,694 | 12,303 | -30% | 1 | 1 | 0% | 3,651 | 2,966 | -19% | 0 | 0 | — |
case-02 | fail→fail | 14,655 | 12,695 | -13% | 1 | 1 | 0% | 3,020 | 3,000 | -1% | 0 | 0 | — |
case-04 | pass→pass | 3,024 | 3,405 | +13% | 1 | 1 | 0% | 690 | 1,283 | +86% | 0 | 0 | — |
case-05 | pass→pass | 3,748 | 3,614 | -4% | 1 | 1 | 0% | 865 | 1,324 | +53% | 0 | 0 | — |
case-06 | pass→pass | 7,781 | 6,111 | -21% | 1 | 1 | 0% | 1,401 | 1,606 | +15% | 0 | 0 | — |
case-07 | pass→pass | 9,882 | 5,738 | -42% | 1 | 1 | 0% | 1,643 | 1,519 | -8% | 0 | 0 | — |
case-08 | pass→pass | 10,319 | 7,011 | -32% | 1 | 1 | 0% | 1,713 | 1,783 | +4% | 0 | 0 | — |
case-09 | pass→pass | 2,517 | 2,984 | +19% | 1 | 1 | 0% | 503 | 1,147 | +128% | 0 | 0 | — |
case-10 | pass→pass | 6,308 | 8,270 | +31% | 1 | 1 | 0% | 1,436 | 2,290 | +59% | 0 | 0 | — |
case-11 | pass→pass | 3,197 | 4,584 | +43% | 1 | 1 | 0% | 721 | 1,461 | +103% | 0 | 0 | — |
case-12 | pass→pass | 3,755 | 5,126 | +37% | 1 | 1 | 0% | 839 | 1,535 | +83% | 0 | 0 | — |
case-13 | pass→pass | 6,822 | 7,905 | +16% | 1 | 1 | 0% | 1,414 | 2,230 | +58% | 0 | 0 | — |
case-14 | pass→pass | 10,603 | 9,522 | -10% | 1 | 1 | 0% | 2,010 | 2,575 | +28% | 0 | 0 | — |
case-15 | pass→pass | 10,806 | 10,249 | -5% | 1 | 1 | 0% | 2,128 | 2,426 | +14% | 0 | 0 | — |
case-16 | pass→pass | 5,403 | 5,242 | -3% | 1 | 1 | 0% | 1,065 | 1,551 | +46% | 0 | 0 | — |
case-17 | pass→pass | 3,086 | 4,380 | +42% | 1 | 1 | 0% | 561 | 1,310 | +134% | 0 | 0 | — |
case-18 | pass→pass | 4,634 | 5,066 | +9% | 1 | 1 | 0% | 914 | 1,596 | +75% | 0 | 0 | — |
case-19 | pass→pass | 10,699 | 12,563 | +17% | 1 | 1 | 0% | 1,670 | 2,653 | +59% | 0 | 0 | — |
case-20 | pass→fail | 18,592 | 17,057 | -8% | 1 | 1 | 0% | 3,218 | 3,481 | +8% | 0 | 0 | — |
case-21 | pass→pass | 7,559 | 9,651 | +28% | 1 | 1 | 0% | 1,577 | 2,567 | +63% | 0 | 0 | — |
case-22 | pass→pass | 24,801 | 24,959 | +1% | 1 | 1 | 0% | 6,166 | 6,729 | +9% | 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. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.