Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Наложить пронумерованную сетку на изображение. Используй при анализе скриншота макета или печатной формы — измерить пропорции колонок перед генерацией табличного документа
.claude/skills/nikolay-shirokov-img-grid/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 324% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -29% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -26% | 0% |
Накладывает пронумерованную сетку на изображение печатной формы. Позволяет точно определить границы колонок, их пропорции и span-ы для генерации макета табличного документа.
/img-grid <ImagePath> [-c COLS] [-o OUTPUT]| Параметр | Обязательный | По умолчанию | Описание | |-----------|:------------:|--------------|-----------------------------------------------| | ImagePath | да | — | Путь к изображению (PNG, JPG) | | -c COLS | нет | 50 | Количество вертикальных делений | | -r ROWS | нет | авто | Количество горизонтальных делений (авто = квадратные ячейки) | | -o OUTPUT | нет | <name>-grid.<ext> | Путь для результата |
bashpython "${CLAUDE_SKILL_DIR}/scripts/overlay-grid.py" "<ImagePath>" [-c 50] [-o "<OutputPath>"]
Требуется Python 3 с библиотекой Pillow (pip install Pillow).
Посмотреть на изображение с сеткой и записать координаты вертикальных границ каждой колонки таблицы (в номерах grid-линий).
Если на форме несколько таблиц с разной раскладкой (например, шапка документа и основная таблица), объединить все граничные точки. Каждый сегмент между соседними границами — одна колонка MXL.
Пример для М-11:
json{ "columns": 16, "page": "A4-landscape", "columnWidths": { "1": "2x", "2": "2x", "3": "5x", "4": "2x", "5": "3x", "6": "2x", "7": "3x", "8": "2x", "9": "2x", "10": "5x", "11": "4x", "12": "2x", "13": "2x", "14": "4x", "15": "2x", "16": "6x" } }
Поле "page" позволяет компилятору автоматически вычислить абсолютные ширины из пропорций.
/mxl-compile → /mxl-validate → /mxl-info
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 1,774 | 4,920 | +177% | 1 | 1 | 0% | 298 | 1,264 | +324% | 0 | 0 | — |
case-02 | fail→pass | 10,203 | 1,943 | -81% | 1 | 1 | 0% | 1,748 | 1,248 | -29% | 0 | 0 | — |
case-03 | fail→pass | 14,773 | 1,729 | -88% | 1 | 1 | 0% | 3,215 | 1,202 | -63% | 0 | 0 | — |
case-04 | fail→pass | 34,617 | 3,151 | -91% | 1 | 1 | 0% | 2,156 | 1,204 | -44% | 0 | 0 | — |
case-05 | fail→pass | 13,532 | 3,787 | -72% | 1 | 1 | 0% | 2,018 | 1,502 | -26% | 0 | 0 | — |
case-06 | pass→pass | 4,924 | 2,113 | -57% | 1 | 1 | 0% | 697 | 1,214 | +74% | 0 | 0 | — |
case-07 | pass→pass | 3,815 | 2,079 | -46% | 1 | 1 | 0% | 498 | 1,203 | +142% | 0 | 0 | — |
case-08 | pass→pass | 8,493 | 2,207 | -74% | 1 | 1 | 0% | 1,288 | 1,223 | -5% | 0 | 0 | — |
case-09 | pass→pass | 4,174 | 4,093 | -2% | 1 | 1 | 0% | 726 | 1,769 | +144% | 0 | 0 | — |
case-10 | pass→pass | 7,989 | 1,983 | -75% | 1 | 1 | 0% | 1,221 | 1,195 | -2% | 0 | 0 | — |
case-11 | pass→pass | 13,941 | 3,507 | -75% | 1 | 1 | 0% | 2,464 | 1,535 | -38% | 0 | 0 | — |
case-12 | fail→pass | 11,866 | 4,267 | -64% | 1 | 1 | 0% | 1,940 | 1,712 | -12% | 0 | 0 | — |
case-13 | fail→pass | 16,820 | 2,331 | -86% | 1 | 1 | 0% | 2,864 | 1,286 | -55% | 0 | 0 | — |
case-14 | fail→pass | 10,867 | 2,589 | -76% | 1 | 1 | 0% | 1,753 | 1,326 | -24% | 0 | 0 | — |
case-15 | pass→pass | 9,534 | 2,082 | -78% | 1 | 1 | 0% | 1,473 | 1,205 | -18% | 0 | 0 | — |
case-16 | pass→pass | 14,626 | 3,360 | -77% | 1 | 1 | 0% | 2,687 | 1,485 | -45% | 0 | 0 | — |
case-17 | pass→pass | 3,564 | 1,476 | -59% | 1 | 1 | 0% | 473 | 1,098 | +132% | 0 | 0 | — |
case-18 | fail→pass | 19,571 | 2,766 | -86% | 1 | 1 | 0% | 1,407 | 1,358 | -3% | 0 | 0 | — |
case-19 | pass→pass | 9,813 | 2,145 | -78% | 1 | 1 | 0% | 1,428 | 1,177 | -18% | 0 | 0 | — |
case-20 | pass→pass | 33,236 | 4,188 | -87% | 1 | 1 | 0% | 2,205 | 1,610 | -27% | 0 | 0 | — |
case-21 | fail→pass | 8,377 | 8,682 | +4% | 1 | 1 | 0% | 488 | 2,318 | +375% | 0 | 0 | — |
case-22 | pass→pass | 4,134 | 13,544 | +228% | 1 | 1 | 0% | 582 | 2,096 | +260% | 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 19 counted toward the lift figure. The other 3 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 +45 percentage points is the difference between those two pass rates over the 19 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.