Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →TSV data processing — view, filter, convert tab-separated data. Zero barrier — use with x-cmd, awk, or Python. Use for "tsv", "tab-separated", "data", "conversion".
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -80% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -61% | 0% |
Process TSV (Tab-Separated Values) data: view, filter, convert between formats.
bash# With x-cmd x tsv < data.tsv # Display TSV data x tsv --csv < data.tsv # Convert to CSV x tsv awk '{print $1, $3}' < data.tsv # Process with AWK # Without x-cmd — AWK handles TSV natively awk -F'\t' '{print $1, $3}' data.tsv # Convert TSV to CSV awk -F'\t' -v OFS=',' '{$1=$1; print}' data.tsv # Filter rows awk -F'\t' '$3 > 100 {print}' data.tsv
x tsv — Display TSV datax tsv --csv — Convert to CSVx tsv awk '{...}' — AWK-style processingawk -F'\t' '{...}' — native TSV supportcsv module with delimiter='\t'mlr --tsv cat data.tsvStarting 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 | 41,974 | 4,472 | -89% | 1 | 1 | 0% | 1,263 | 868 | -31% | 0 | 0 | — |
case-02 | fail→pass | 70,565 | 3,072 | -96% | 1 | 1 | 0% | 1,150 | 727 | -37% | 0 | 0 | — |
case-03 | fail→pass | 56,604 | 4,344 | -92% | 1 | 1 | 0% | 3,655 | 715 | -80% | 0 | 0 | — |
case-04 | pass→pass | 49,588 | 8,536 | -83% | 1 | 1 | 0% | 2,790 | 821 | -71% | 0 | 0 | — |
case-05 | fail→pass | 32,744 | 15,629 | -52% | 1 | 1 | 0% | 2,131 | 1,275 | -40% | 0 | 0 | — |
case-06 | pass→pass | 36,738 | 3,730 | -90% | 1 | 1 | 0% | 1,340 | 876 | -35% | 0 | 0 | — |
case-07 | fail→pass | 18,458 | 4,806 | -74% | 1 | 1 | 0% | 2,715 | 1,072 | -61% | 0 | 0 | — |
case-08 | fail→pass | 9,792 | 4,630 | -53% | 1 | 1 | 0% | 1,544 | 953 | -38% | 0 | 0 | — |
case-09 | fail→fail | 18,324 | 5,385 | -71% | 1 | 1 | 0% | 1,959 | 1,206 | -38% | 0 | 0 | — |
case-10 | pass→pass | 8,089 | 7,948 | -2% | 1 | 1 | 0% | 1,194 | 1,003 | -16% | 0 | 0 | — |
case-11 | fail→pass | 21,690 | 3,246 | -85% | 1 | 1 | 0% | 1,292 | 789 | -39% | 0 | 0 | — |
case-12 | fail→pass | 10,859 | 129,929 | +1097% | 1 | 1 | 0% | 1,737 | 895 | -48% | 0 | 0 | — |
case-13 | fail→pass | 13,405 | 5,474 | -59% | 1 | 1 | 0% | 1,673 | 974 | -42% | 0 | 0 | — |
case-14 | fail→fail | 14,367 | 5,282 | -63% | 1 | 1 | 0% | 1,333 | 1,153 | -14% | 0 | 0 | — |
case-15 | fail→pass | 8,348 | 5,389 | -35% | 1 | 1 | 0% | 1,363 | 1,062 | -22% | 0 | 0 | — |
case-16 | fail→pass | 38,736 | 12,771 | -67% | 1 | 1 | 0% | 1,716 | 1,031 | -40% | 0 | 0 | — |
case-17 | fail→pass | 9,217 | 4,840 | -47% | 1 | 1 | 0% | 1,336 | 911 | -32% | 0 | 0 | — |
case-18 | fail→pass | 15,125 | 5,454 | -64% | 1 | 1 | 0% | 1,302 | 876 | -33% | 0 | 0 | — |
case-19 | fail→pass | 7,405 | 5,544 | -25% | 1 | 1 | 0% | 1,082 | 1,062 | -2% | 0 | 0 | — |
case-20 | fail→pass | 46,090 | 8,344 | -82% | 1 | 1 | 0% | 2,582 | 1,771 | -31% | 0 | 0 | — |
case-21 | pass→pass | 9,902 | 7,535 | -24% | 1 | 1 | 0% | 1,504 | 1,470 | -2% | 0 | 0 | — |
case-22 | fail→pass | 10,435 | 8,535 | -18% | 1 | 1 | 0% | 1,552 | 1,128 | -27% | 0 | 0 | — |
case-23 | pass→fail | 5,016 | 11,017 | +120% | 1 | 1 | 0% | 683 | 988 | +45% | 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 +65 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.