Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Chinese Calendar with Lunar-Solar Conversion
.claude/skills/x-cmd-ccal/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -67% | 0% |
调休 — the rule that a weekend may be borrowed to make a weekday-off, or a weekday given back when a holiday falls on a weekend — is the key feature for Chinese users planning around the official work calendar.
x ccalbash# Show all date info in tsv form. TSV column as below: # Date LunarDate 建除 WeekDay(周几) Jieqi(节气) Xiuxi(三个状态:休，工，无-非调休导致的假期或工作日) Yi（宜） Ji（忌） Holiday x ccal ls 2026 # Show all date info in tsv form x ccal ls 2026-07 # Specific date in YAML x ccal info 2026-05-31 # Get data directory. If it is empty, try `x ccal update` x ccal datadir
x ccal -h for more info.
Notice: the data source is https://codeberg.org/x-cmd/ccal-data/releases/download/latest/ccal-data.tar.xz
To save space, x ccal navigate the tar.xz using command tar without decompression. 我们鼓励这种方法，但如果需要更自由的处理，你可以按需解压。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 14,566 | 3,983 | -73% | 1 | 1 | 0% | 2,493 | 910 | -63% | 0 | 0 | — |
case-02 | fail→pass | 8,777 | 2,663 | -70% | 1 | 1 | 0% | 1,326 | 629 | -53% | 0 | 0 | — |
case-03 | fail→pass | 9,271 | 2,844 | -69% | 1 | 1 | 0% | 1,404 | 595 | -58% | 0 | 0 | — |
case-04 | pass→pass | 7,116 | 4,724 | -34% | 1 | 1 | 0% | 1,046 | 1,052 | +1% | 0 | 0 | — |
case-05 | pass→fail | 5,627 | 5,127 | -9% | 1 | 1 | 0% | 1,063 | 976 | -8% | 0 | 0 | — |
case-06 | pass→pass | 10,516 | 11,303 | +7% | 1 | 1 | 0% | 2,030 | 2,277 | +12% | 0 | 0 | — |
case-07 | fail→pass | 13,129 | 6,392 | -51% | 1 | 1 | 0% | 2,350 | 791 | -66% | 0 | 0 | — |
case-08 | pass→pass | 15,464 | 2,615 | -83% | 1 | 1 | 0% | 2,121 | 642 | -70% | 0 | 0 | — |
case-09 | pass→pass | 14,841 | 6,259 | -58% | 1 | 1 | 0% | 2,033 | 1,244 | -39% | 0 | 0 | — |
case-10 | fail→pass | 21,885 | 2,720 | -88% | 1 | 1 | 0% | 2,178 | 723 | -67% | 0 | 0 | — |
case-11 | fail→pass | 31,679 | 4,603 | -85% | 1 | 1 | 0% | 2,115 | 1,072 | -49% | 0 | 0 | — |
case-12 | pass→pass | 14,140 | 7,943 | -44% | 1 | 1 | 0% | 831 | 1,514 | +82% | 0 | 0 | — |
case-13 | fail→pass | 15,243 | 3,734 | -76% | 1 | 1 | 0% | 1,461 | 867 | -41% | 0 | 0 | — |
case-14 | fail→pass | 19,930 | 3,482 | -83% | 1 | 1 | 0% | 2,059 | 838 | -59% | 0 | 0 | — |
case-15 | pass→pass | 8,245 | 3,938 | -52% | 1 | 1 | 0% | 1,436 | 901 | -37% | 0 | 0 | — |
case-16 | fail→pass | 11,473 | 3,048 | -73% | 1 | 1 | 0% | 1,716 | 707 | -59% | 0 | 0 | — |
case-17 | fail→pass | 12,832 | 7,056 | -45% | 1 | 1 | 0% | 2,185 | 935 | -57% | 0 | 0 | — |
case-18 | pass→pass | 4,006 | 2,651 | -34% | 1 | 1 | 0% | 576 | 511 | -11% | 0 | 0 | — |
case-19 | pass→pass | 9,906 | 5,740 | -42% | 1 | 1 | 0% | 1,691 | 1,143 | -32% | 0 | 0 | — |
case-20 | pass→pass | 9,204 | 4,782 | -48% | 1 | 1 | 0% | 1,512 | 823 | -46% | 0 | 0 | — |
case-21 | pass→pass | 9,437 | 3,647 | -61% | 1 | 1 | 0% | 1,476 | 852 | -42% | 0 | 0 | — |
case-22 | fail→pass | 11,056 | 2,978 | -73% | 1 | 1 | 0% | 2,190 | 714 | -67% | 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 +45 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.
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.