Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create new Warp tab config TOML files from natural-language requests. Use when the user wants a new tab config, a new tab layout, or asks for a slash command to generate a tab config.
.claude/skills/warpdotdev-create-tab-config/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✓→✓ | ＝ Same ✓ | 69% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 100% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 20% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -70% | 0% |
| case-02 | ✗→✗ | ＝ Same ✗ | -55% | 0% |
Create a new Warp tab config based on what the user wants.
tab-configs skill as the canonical source of truth for:ask_user_question tool to clarify them before writing anything. Do not guess about layout, commands, directories, parameters, or close-time behavior.tab-configs schema.Tab configs live under ~/.warp/tab_configs/ for standard builds. Non-stable builds use a channel-specific variant (e.g. ~/.warp-<channel>/tab_configs/). To find the correct directory, run ls -d ~/.warp*/ to list the available Warp data directories and pick the one that corresponds to the running build. When in doubt, ask the user which build they are using. Create the tab_configs/ subdirectory if it does not exist. Write the file using a descriptive snake_case filename ending in .toml.
ask_user_question tool.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 12,836 | 25,188 | +96% | 1 | 1 | 0% | 2,117 | 626 | -70% | 0 | 0 | — |
case-02 | fail→fail | 8,110 | 7,439 | -8% | 1 | 1 | 0% | 1,559 | 705 | -55% | 0 | 0 | — |
case-03 | fail→fail | 10,116 | 16,509 | +63% | 1 | 1 | 0% | 1,824 | 1,587 | -13% | 0 | 0 | — |
case-04 | pass→pass | 12,043 | 14,875 | +24% | 1 | 1 | 0% | 1,852 | 3,139 | +69% | 0 | 0 | — |
case-05 | pass→pass | 11,200 | 19,862 | +77% | 1 | 1 | 0% | 2,138 | 4,280 | +100% | 0 | 0 | — |
case-06 | pass→pass | 11,590 | 12,374 | +7% | 1 | 1 | 0% | 1,976 | 2,373 | +20% | 0 | 0 | — |
case-07 | fail→fail | 11,684 | 8,297 | -29% | 1 | 1 | 0% | 2,010 | 839 | -58% | 0 | 0 | — |
case-08 | fail→fail | 12,264 | 17,042 | +39% | 1 | 1 | 0% | 1,947 | 1,710 | -12% | 0 | 0 | — |
case-09 | fail→fail | 11,330 | 43,101 | +280% | 1 | 1 | 0% | 1,970 | 8,417 | +327% | 0 | 0 | — |
case-10 | fail→fail | 8,716 | 10,016 | +15% | 1 | 1 | 0% | 1,507 | 705 | -53% | 0 | 0 | — |
case-11 | fail→fail | 12,911 | 37,004 | +187% | 1 | 1 | 0% | 2,279 | 1,293 | -43% | 0 | 0 | — |
case-12 | fail→fail | 9,215 | 8,466 | -8% | 1 | 1 | 0% | 1,640 | 734 | -55% | 0 | 0 | — |
case-13 | fail→fail | 23,451 | 19,675 | -16% | 1 | 1 | 0% | 2,433 | 2,835 | +17% | 0 | 0 | — |
case-14 | fail→fail | 18,907 | 16,018 | -15% | 1 | 1 | 0% | 3,488 | 766 | -78% | 0 | 0 | — |
case-15 | fail→fail | 11,954 | 10,296 | -14% | 1 | 1 | 0% | 2,345 | 1,038 | -56% | 0 | 0 | — |
case-16 | fail→fail | 7,499 | 20,130 | +168% | 1 | 1 | 0% | 1,209 | 1,632 | +35% | 0 | 0 | — |
case-17 | fail→fail | 15,358 | 7,563 | -51% | 1 | 1 | 0% | 2,758 | 641 | -77% | 0 | 0 | — |
case-18 | fail→fail | 10,838 | 13,499 | +25% | 1 | 1 | 0% | 1,606 | 712 | -56% | 0 | 0 | — |
case-19 | fail→fail | 10,739 | 7,312 | -32% | 1 | 1 | 0% | 1,673 | 667 | -60% | 0 | 0 | — |
case-20 | fail→fail | 9,553 | 7,556 | -21% | 1 | 1 | 0% | 1,716 | 842 | -51% | 0 | 0 | — |
case-21 | fail→fail | 8,589 | 15,907 | +85% | 1 | 1 | 0% | 1,507 | 1,269 | -16% | 0 | 0 | — |
case-22 | fail→fail | 9,693 | 10,171 | +5% | 1 | 1 | 0% | 1,549 | 757 | -51% | 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 6 counted toward the lift figure. The other 16 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 0 percentage points is the difference between those two pass rates over the 6 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Other measured skills in the registry, with their headline benchmark lift.