Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →How to add i18n to all the ts files, and use the compile_i18n.cmd command to compile the ts files
.claude/skills/compile-i18n/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -75% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -40% | 0% |
When you are instructed to add or update internationalization (i18n) translations for this project, please follow these steps:
.ts translation files located in the ok/ui/qt/i18n/ directory. Typically, these files include:zh_CN.ts (Simplified Chinese)zh_TW.ts (Traditional Chinese)ja_JP.ts (Japanese)ko_KR.ts (Korean)es_ES.ts (Spanish)en_US.ts (English)<context> and <message> blocks in each XML file, without the type="unfinished" attribute. If you are adding many translations, or just want to minimize errors, use the provided add_translation.py script to automate the XML injection:powershell $py = if (Test-Path .\.venv\Scripts\python.exe) { ".\.venv\Scripts\python.exe" } else { "python" } & $py .\.agents\skills\compile_i18n\add_translation.py --context "YourContext" --source "Your Source String" --zh_cn "Your Chinese String" IMPORTANT WARNING: DO NOT use pyside6-lupdate as it corrupts and deletes existing translations. Always modify the XML nodes natively or use the python script!
WebUI context are complete. You must replace <translation type="unfinished" /> with <translation>Your Translation</translation>.WebUI context by running this command from the workspace root:// turbo
powershellnpm run i18n
npm run dev and npm run build also run this generator automatically.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,969 | 6,708 | -16% | 1 | 1 | 0% | 1,445 | 730 | -49% | 0 | 0 | — |
case-02 | fail→fail | 9,083 | 27,049 | +198% | 1 | 1 | 0% | 1,397 | 694 | -50% | 0 | 0 | — |
case-03 | fail→fail | 6,614 | 7,111 | +8% | 1 | 1 | 0% | 950 | 1,008 | +6% | 0 | 0 | — |
case-04 | fail→pass | 10,351 | 4,301 | -58% | 1 | 1 | 0% | 1,943 | 1,296 | -33% | 0 | 0 | — |
case-05 | pass→pass | 13,523 | 3,778 | -72% | 1 | 1 | 0% | 2,331 | 1,072 | -54% | 0 | 0 | — |
case-06 | fail→pass | 10,534 | 3,223 | -69% | 1 | 1 | 0% | 1,819 | 1,070 | -41% | 0 | 0 | — |
case-07 | fail→pass | 10,240 | 2,123 | -79% | 1 | 1 | 0% | 1,763 | 740 | -58% | 0 | 0 | — |
case-08 | pass→pass | 9,007 | 2,385 | -74% | 1 | 1 | 0% | 1,654 | 861 | -48% | 0 | 0 | — |
case-09 | fail→pass | 13,886 | 1,249 | -91% | 1 | 1 | 0% | 2,545 | 626 | -75% | 0 | 0 | — |
case-10 | fail→pass | 7,634 | 2,305 | -70% | 1 | 1 | 0% | 1,331 | 803 | -40% | 0 | 0 | — |
case-11 | fail→pass | 8,308 | 3,103 | -63% | 1 | 1 | 0% | 1,352 | 963 | -29% | 0 | 0 | — |
case-24 | pass→pass | 10,647 | 7,741 | -27% | 1 | 1 | 0% | 2,204 | 2,044 | -7% | 0 | 0 | — |
case-12 | pass→pass | 10,520 | 1,823 | -83% | 1 | 1 | 0% | 1,680 | 659 | -61% | 0 | 0 | — |
case-13 | pass→pass | 6,226 | 1,571 | -75% | 1 | 1 | 0% | 1,129 | 689 | -39% | 0 | 0 | — |
case-14 | fail→pass | 9,345 | 3,454 | -63% | 1 | 1 | 0% | 1,500 | 1,145 | -24% | 0 | 0 | — |
case-15 | fail→pass | 7,275 | 1,621 | -78% | 1 | 1 | 0% | 1,241 | 605 | -51% | 0 | 0 | — |
case-16 | pass→pass | 5,045 | 2,166 | -57% | 1 | 1 | 0% | 858 | 737 | -14% | 0 | 0 | — |
case-17 | pass→pass | 10,122 | 1,679 | -83% | 1 | 1 | 0% | 1,548 | 686 | -56% | 0 | 0 | — |
case-18 | pass→pass | 11,654 | 1,860 | -84% | 1 | 1 | 0% | 1,898 | 719 | -62% | 0 | 0 | — |
case-19 | fail→pass | 7,490 | 3,060 | -59% | 1 | 1 | 0% | 1,286 | 1,021 | -21% | 0 | 0 | — |
case-20 | fail→fail | 7,424 | 1,773 | -76% | 1 | 1 | 0% | 1,363 | 738 | -46% | 0 | 0 | — |
case-21 | pass→pass | 8,905 | 1,816 | -80% | 1 | 1 | 0% | 1,369 | 743 | -46% | 0 | 0 | — |
case-22 | fail→pass | 11,417 | 2,511 | -78% | 1 | 1 | 0% | 2,022 | 880 | -56% | 0 | 0 | — |
case-23 | pass→pass | 10,540 | 3,793 | -64% | 1 | 1 | 0% | 1,968 | 980 | -50% | 0 | 0 | — |
case-25 | pass→pass | 3,912 | 3,901 | -0% | 1 | 1 | 0% | 570 | 1,091 | +91% | 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. 25 cases were attempted, and 23 counted toward the lift figure. The other 2 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 +40 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/28/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.