Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →CSV data processing — view, filter, convert, merge tabular data. Zero barrier — use with x-cmd, Python, or awk. Use for "csv", "data", "table", "spreadsheet", "conversion".
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -20% | 0% |
Process CSV data: view, filter, convert, merge. Works with any tabular data source.
bash# With x-cmd x csv tab 1,3,6 < data.csv # Show specific columns x csv awk '{print $1, $3}' < data.csv # Process with AWK x csv tojson < data.csv # Convert to JSON x csv merge2 id file2.csv < file1.csv # Merge by key # Without x-cmd — use Python python3 -c " import csv, json, sys reader = csv.DictReader(sys.stdin) for row in reader: print(row['name'], row['value']) " < data.csv # Or use Miller (mlr), xsv, csvkit
x csv tab <cols> — Display selected columnsx csv awk '{...}' — AWK-style processingx csv tojson — Convert to JSONx csv tojsonl — Convert to JSON Linesx csv totsv — Convert to TSVx csv merge2 — Merge two CSVs by keyx csv header — Operate on headersx csv app — Interactive table viewercsv module, pandasmlr (Miller), xsv, csvcut, csvgrepawk -F, '{print $1}'Starting with the essentials. Will add:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 22,652 | 2,956 | -87% | 1 | 1 | 0% | 923 | 650 | -30% | 0 | 0 | — |
case-02 | fail→pass | 9,200 | 4,376 | -52% | 1 | 1 | 0% | 1,285 | 865 | -33% | 0 | 0 | — |
case-03 | fail→pass | 75,089 | 4,117 | -95% | 1 | 1 | 0% | 1,613 | 819 | -49% | 0 | 0 | — |
case-04 | fail→pass | 60,781 | 5,058 | -92% | 1 | 1 | 0% | 1,091 | 786 | -28% | 0 | 0 | — |
case-05 | fail→pass | 19,048 | 3,571 | -81% | 1 | 1 | 0% | 1,011 | 811 | -20% | 0 | 0 | — |
case-06 | pass→pass | 8,666 | 4,189 | -52% | 1 | 1 | 0% | 1,113 | 763 | -31% | 0 | 0 | — |
case-07 | pass→pass | 8,332 | 3,061 | -63% | 1 | 1 | 0% | 918 | 746 | -19% | 0 | 0 | — |
case-08 | pass→pass | 13,979 | 5,502 | -61% | 1 | 1 | 0% | 1,879 | 1,023 | -46% | 0 | 0 | — |
case-09 | fail→pass | 9,781 | 6,027 | -38% | 1 | 1 | 0% | 1,610 | 927 | -42% | 0 | 0 | — |
case-10 | pass→pass | 10,363 | 12,167 | +17% | 1 | 1 | 0% | 1,618 | 1,196 | -26% | 0 | 0 | — |
case-11 | pass→pass | 16,647 | 5,700 | -66% | 1 | 1 | 0% | 1,197 | 937 | -22% | 0 | 0 | — |
case-12 | pass→pass | 10,698 | 3,341 | -69% | 1 | 1 | 0% | 1,455 | 831 | -43% | 0 | 0 | — |
case-13 | pass→pass | 4,640 | 4,553 | -2% | 1 | 1 | 0% | 390 | 806 | +107% | 0 | 0 | — |
case-14 | pass→pass | 3,100 | 6,328 | +104% | 1 | 1 | 0% | 336 | 778 | +132% | 0 | 0 | — |
case-15 | pass→pass | 3,850 | 4,363 | +13% | 1 | 1 | 0% | 465 | 915 | +97% | 0 | 0 | — |
case-16 | pass→pass | 33,328 | 6,979 | -79% | 1 | 1 | 0% | 1,919 | 1,253 | -35% | 0 | 0 | — |
case-17 | fail→pass | 8,630 | 2,717 | -69% | 1 | 1 | 0% | 1,249 | 693 | -45% | 0 | 0 | — |
case-18 | fail→pass | 10,161 | 3,332 | -67% | 1 | 1 | 0% | 1,569 | 689 | -56% | 0 | 0 | — |
case-19 | fail→pass | 12,545 | 2,125 | -83% | 1 | 1 | 0% | 1,873 | 638 | -66% | 0 | 0 | — |
case-20 | pass→pass | 18,525 | 6,417 | -65% | 1 | 1 | 0% | 1,393 | 1,311 | -6% | 0 | 0 | — |
case-21 | pass→pass | 21,366 | 9,847 | -54% | 1 | 1 | 0% | 1,896 | 2,051 | +8% | 0 | 0 | — |
case-22 | pass→pass | 4,658 | 7,928 | +70% | 1 | 1 | 0% | 615 | 932 | +52% | 0 | 0 | — |
case-23 | pass→pass | 8,463 | 3,797 | -55% | 1 | 1 | 0% | 1,030 | 811 | -21% | 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 +39 percentage points is the difference between those two pass rates over the 23 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.