Auto-build: copilot (powershell) from 113dc9e

.github/skills/mxl-validate/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: mxl-validate
+description: Валидация макета табличного документа (MXL). Используй после создания или модификации макета для проверки корректности
+argument-hint: <TemplatePath> [-Detailed] [-MaxErrors 20]
+allowed-tools:
+  - Bash
+  - Read
+  - Glob
+---
+
+# /mxl-validate — валидация макета табличного документа (MXL)
+
+Проверяет Template.xml на структурные ошибки: индексы, ссылки на палитры, диапазоны именованных областей и объединений, согласованность ячеек-полей ввода.
+
+## Параметры
+
+| Параметр      | Обяз. | Умолч. | Описание                                 |
+|---------------|:-----:|---------|--------------------------------------------|
+| TemplatePath  | да    | —       | Путь к макету (директория или Template.xml) |
+| Detailed      | нет   | —       | Подробный вывод (все проверки, включая успешные) |
+| MaxErrors     | нет   | 20      | Остановиться после N ошибок                |
+
+## Команда
+
+```powershell
+powershell.exe -NoProfile -File ".github/skills/mxl-validate/scripts/mxl-validate.ps1" -TemplatePath "Catalogs/Номенклатура/Templates/Макет"
+powershell.exe -NoProfile -File ".github/skills/mxl-validate/scripts/mxl-validate.ps1" -TemplatePath "src/МояОбработка/Templates/ПечатнаяФорма"
+```
+

.github/skills/mxl-validate/scripts/mxl-validate.ps1

@@ -0,0 +1,572 @@
+# mxl-validate v1.7 — Validate 1C spreadsheet
+# Source: https://github.com/Nikolay-Shirokov/cc-1c-skills
+param(
+	[Alias('Path')]
+	[string]$TemplatePath,
+	[string]$ProcessorName,
+	[string]$TemplateName,
+	[string]$SrcDir = "src",
+	[switch]$Detailed,
+	[int]$MaxErrors = 20
+)
+
+$ErrorActionPreference = "Stop"
+[Console]::OutputEncoding = [System.Text.Encoding]::UTF8
+
+# --- Resolve template path ---
+
+if (-not $TemplatePath) {
+	if (-not $ProcessorName -or -not $TemplateName) {
+		Write-Error "Specify -TemplatePath or both -ProcessorName and -TemplateName"
+		exit 1
+	}
+	$TemplatePath = Join-Path (Join-Path (Join-Path (Join-Path (Join-Path $SrcDir $ProcessorName) "Templates") $TemplateName) "Ext") "Template.xml"
+}
+
+# A: Directory → Ext/Template.xml
+if (Test-Path $TemplatePath -PathType Container) {
+	$TemplatePath = Join-Path (Join-Path $TemplatePath "Ext") "Template.xml"
+}
+# B1: Missing Ext/ (e.g. Templates/Макет/Template.xml → Templates/Макет/Ext/Template.xml)
+if (-not (Test-Path $TemplatePath)) {
+	$fn = [System.IO.Path]::GetFileName($TemplatePath)
+	if ($fn -eq "Template.xml") {
+		$c = Join-Path (Join-Path (Split-Path $TemplatePath) "Ext") $fn
+		if (Test-Path $c) { $TemplatePath = $c }
+	}
+}
+# B2: Descriptor (Templates/Макет.xml → Templates/Макет/Ext/Template.xml)
+if (-not (Test-Path $TemplatePath) -and $TemplatePath.EndsWith(".xml")) {
+	$stem = [System.IO.Path]::GetFileNameWithoutExtension($TemplatePath)
+	$dir = Split-Path $TemplatePath
+	$c = Join-Path (Join-Path (Join-Path $dir $stem) "Ext") "Template.xml"
+	if (Test-Path $c) { $TemplatePath = $c }
+}
+
+if (-not (Test-Path $TemplatePath)) {
+	Write-Error "File not found: $TemplatePath"
+	exit 1
+}
+
+# --- Load XML ---
+
+$xmlDoc = New-Object System.Xml.XmlDocument
+$xmlDoc.PreserveWhitespace = $false
+$xmlDoc.Load((Resolve-Path $TemplatePath).Path)
+
+$nsMgr = New-Object System.Xml.XmlNamespaceManager($xmlDoc.NameTable)
+$nsMgr.AddNamespace("d", "http://v8.1c.ru/8.2/data/spreadsheet")
+$nsMgr.AddNamespace("v8", "http://v8.1c.ru/8.1/data/core")
+$nsMgr.AddNamespace("xsi", "http://www.w3.org/2001/XMLSchema-instance")
+
+$root = $xmlDoc.DocumentElement
+
+# --- Counters ---
+
+$errors = 0
+$warnings = 0
+$stopped = $false
+$script:okCount = 0
+
+function Report-OK {
+	param([string]$msg)
+	$script:okCount++
+	if ($Detailed) { Write-Host "[OK]    $msg" }
+}
+
+function Report-Error {
+	param([string]$msg)
+	$script:errors++
+	Write-Host "[ERROR] $msg"
+	if ($script:errors -ge $MaxErrors) {
+		$script:stopped = $true
+	}
+}
+
+function Report-Warn {
+	param([string]$msg)
+	$script:warnings++
+	Write-Host "[WARN]  $msg"
+}
+
+$templateName = [System.IO.Path]::GetFileName([System.IO.Path]::GetDirectoryName([System.IO.Path]::GetDirectoryName($TemplatePath)))
+if ($Detailed) {
+	# Форма имени та же, что в итоговой строке и у py-порта: раньше шапка -Detailed
+	# печаталась без префикса «Template.», и вывод портов расходился одной строкой.
+	Write-Host "=== Validation: Template.$templateName ==="
+	Write-Host ""
+}
+
+# --- Collect palettes ---
+
+$lineNodes = $root.SelectNodes("d:line", $nsMgr)
+$lineCount = $lineNodes.Count
+
+$fontNodes = @()
+foreach ($node in $root.ChildNodes) {
+	if ($node.LocalName -eq "font") { $fontNodes += $node }
+}
+$fontCount = $fontNodes.Count
+
+$formatNodes = @()
+foreach ($node in $root.ChildNodes) {
+	if ($node.LocalName -eq "format") { $formatNodes += $node }
+}
+$formatCount = $formatNodes.Count
+
+$pictureNodes = $root.SelectNodes("d:picture", $nsMgr)
+$pictureCount = $pictureNodes.Count
+
+# --- Collect column sets ---
+
+$columnSets = @{}  # id -> size
+$defaultColCount = 0
+
+foreach ($cols in $root.SelectNodes("d:columns", $nsMgr)) {
+	$sizeNode = $cols.SelectSingleNode("d:size", $nsMgr)
+	$idNode = $cols.SelectSingleNode("d:id", $nsMgr)
+	$size = if ($sizeNode) { [int]$sizeNode.InnerText } else { 0 }
+
+	if ($idNode) {
+		$columnSets[$idNode.InnerText] = $size
+	} else {
+		$defaultColCount = $size
+	}
+}
+
+# --- Check 1: height vs actual rows ---
+
+$rowNodes = $root.SelectNodes("d:rowsItem", $nsMgr)
+$heightNode = $root.SelectSingleNode("d:height", $nsMgr)
+$docHeight = if ($heightNode) { [int]$heightNode.InnerText } else { 0 }
+
+# Find max row index (not all rows have rowsItem - implicit rows are skipped)
+$maxRowIndex = -1
+foreach ($ri in $rowNodes) {
+	$idxNode = $ri.SelectSingleNode("d:index", $nsMgr)
+	if ($idxNode) {
+		$idx = [int]$idxNode.InnerText
+		if ($idx -gt $maxRowIndex) { $maxRowIndex = $idx }
+	}
+}
+
+$expectedMinHeight = $maxRowIndex + 1
+if ($docHeight -ge $expectedMinHeight) {
+	Report-OK "height ($docHeight) >= max row index + 1 ($expectedMinHeight), rowsItem count=$($rowNodes.Count)"
+} else {
+	Report-Error "height=$docHeight but max row index=$maxRowIndex (need at least $expectedMinHeight)"
+}
+# --- Check 2: vgRows <= height ---
+
+$vgRowsNode = $root.SelectSingleNode("d:vgRows", $nsMgr)
+if ($vgRowsNode) {
+	$vgRows = [int]$vgRowsNode.InnerText
+	if ($vgRows -le $docHeight) {
+		Report-OK "vgRows ($vgRows) <= height ($docHeight)"
+	} else {
+		Report-Warn "vgRows ($vgRows) > height ($docHeight)"
+	}
+}
+
+# --- Build row data for checks ---
+
+$maxFormatRef = 0
+$maxFontRef = 0
+$maxLineRef = 0
+
+# Check format palette references in formats (font, border indices)
+foreach ($fmt in $formatNodes) {
+	$fontIdx = $fmt.SelectSingleNode("d:font", $nsMgr)
+	if ($fontIdx) {
+		$val = [int]$fontIdx.InnerText
+		if ($val -gt $maxFontRef) { $maxFontRef = $val }
+	}
+
+	foreach ($border in @("d:leftBorder", "d:topBorder", "d:rightBorder", "d:bottomBorder", "d:drawingBorder")) {
+		$borderNode = $fmt.SelectSingleNode($border, $nsMgr)
+		if ($borderNode) {
+			$val = [int]$borderNode.InnerText
+			if ($val -gt $maxLineRef) { $maxLineRef = $val }
+		}
+	}
+}
+
+# --- Check 10: font indices in formats ---
+
+if ($fontCount -gt 0) {
+	if ($maxFontRef -lt $fontCount) {
+		Report-OK "Font refs: max=$maxFontRef, palette size=$fontCount"
+	} else {
+		Report-Error "Font index $maxFontRef exceeds palette size ($fontCount)"
+	}
+} elseif ($maxFontRef -gt 0) {
+	Report-Error "Font index $maxFontRef referenced but no fonts defined"
+} else {
+	Report-OK "No font references"
+}
+
+# --- Check 11: line/border indices in formats ---
+
+if ($lineCount -gt 0) {
+	if ($maxLineRef -lt $lineCount) {
+		Report-OK "Line/border refs: max=$maxLineRef, palette size=$lineCount"
+	} else {
+		Report-Error "Line index $maxLineRef exceeds palette size ($lineCount)"
+	}
+} elseif ($maxLineRef -gt 0) {
+	Report-Error "Line index $maxLineRef referenced but no lines defined"
+} else {
+	Report-OK "No line/border references"
+}
+
+# --- Check 3, 4, 5, 6: row/cell checks ---
+
+$maxCellFormatRef = 0
+$maxRowFormatRef = 0
+$maxDefaultColIdx = 0
+$rowIndex = 0
+
+foreach ($ri in $rowNodes) {
+	if ($stopped) { break }
+
+	$idxNode = $ri.SelectSingleNode("d:index", $nsMgr)
+	$rowIndex = if ($idxNode) { [int]$idxNode.InnerText } else { $rowIndex }
+
+	$row = $ri.SelectSingleNode("d:row", $nsMgr)
+	if (-not $row) { continue }
+
+	# Row formatIndex
+	$rowFmtNode = $row.SelectSingleNode("d:formatIndex", $nsMgr)
+	if ($rowFmtNode) {
+		$val = [int]$rowFmtNode.InnerText
+		if ($val -gt $maxRowFormatRef) { $maxRowFormatRef = $val }
+		if ($val -gt $formatCount) {
+			Report-Error "Row ${rowIndex}: formatIndex=$val > format palette size ($formatCount)"
+		}
+	}
+
+	# Check columnsID
+	$rowColsId = $null
+	$colsIdNode = $row.SelectSingleNode("d:columnsID", $nsMgr)
+	if ($colsIdNode) {
+		$rowColsId = $colsIdNode.InnerText
+		if (-not $columnSets.ContainsKey($rowColsId)) {
+			Report-Error "Row ${rowIndex}: columnsID '$($rowColsId.Substring(0,8))...' not found in column sets"
+		}
+	}
+
+	# Determine column count for this row
+	$rowColCount = $defaultColCount
+	if ($rowColsId -and $columnSets.ContainsKey($rowColsId)) {
+		$rowColCount = $columnSets[$rowColsId]
+	}
+
+	# Cell checks
+	foreach ($cGroup in $row.SelectNodes("d:c", $nsMgr)) {
+		$iNode = $cGroup.SelectSingleNode("d:i", $nsMgr)
+		if ($iNode) {
+			$colIdx = [int]$iNode.InnerText
+			# Track max index for default column set only
+			if (-not $rowColsId -and $colIdx -gt $maxDefaultColIdx) {
+				$maxDefaultColIdx = $colIdx
+			}
+			# Check against row's column count
+			if ($rowColCount -gt 0 -and $colIdx -ge $rowColCount) {
+				Report-Error "Row ${rowIndex}: column index $colIdx >= column count ($rowColCount)"
+			}
+		}
+
+		$cell = $cGroup.SelectSingleNode("d:c", $nsMgr)
+		if ($cell) {
+			$fNode = $cell.SelectSingleNode("d:f", $nsMgr)
+			if ($fNode) {
+				$val = [int]$fNode.InnerText
+				if ($val -gt $maxCellFormatRef) { $maxCellFormatRef = $val }
+				if ($val -gt $formatCount) {
+					Report-Error "Row ${rowIndex}: cell format index $val > format palette size ($formatCount)"
+				}
+			}
+			# Примечание — четвёртый владелец формата, и его ссылка тоже бывает битой.
+			$noteNode = $cell.SelectSingleNode("d:note", $nsMgr)
+			if ($noteNode) {
+				$nf = $noteNode.SelectSingleNode("d:formatIndex", $nsMgr)
+				if ($nf -and [int]$nf.InnerText -gt $formatCount) {
+					Report-Error "Row ${rowIndex}: note format index $($nf.InnerText) > format palette size ($formatCount)"
+				}
+			}
+		}
+	}
+
+	$rowIndex++
+}
+
+# Summary checks for format refs
+if (-not $stopped) {
+	if ($maxCellFormatRef -le $formatCount -and $maxRowFormatRef -le $formatCount) {
+		Report-OK "Format refs: max cell=$maxCellFormatRef, max row=$maxRowFormatRef, palette size=$formatCount"
+	}
+}
+
+# Check column format indices
+foreach ($cols in $root.SelectNodes("d:columns", $nsMgr)) {
+	if ($stopped) { break }
+	foreach ($ci in $cols.SelectNodes("d:columnsItem", $nsMgr)) {
+		$col = $ci.SelectSingleNode("d:column", $nsMgr)
+		if ($col) {
+			$fmtNode = $col.SelectSingleNode("d:formatIndex", $nsMgr)
+			if ($fmtNode) {
+				$val = [int]$fmtNode.InnerText
+				if ($val -gt $formatCount) {
+					$colIdx = $ci.SelectSingleNode("d:index", $nsMgr).InnerText
+					Report-Error "Column ${colIdx}: formatIndex=$val > format palette size ($formatCount)"
+				}
+			}
+		}
+	}
+}
+
+# --- Check 5: column index summary ---
+
+if (-not $stopped) {
+	Report-OK "Column indices: max in default set=$maxDefaultColIdx, default column count=$defaultColCount"
+}
+
+# --- Check 7, 8: named areas ---
+
+foreach ($ni in $root.SelectNodes("d:namedItem", $nsMgr)) {
+	if ($stopped) { break }
+
+	$niType = $ni.GetAttribute("type", "http://www.w3.org/2001/XMLSchema-instance")
+	$name = $ni.SelectSingleNode("d:name", $nsMgr).InnerText
+
+	if ($niType -like "*NamedItemCells*") {
+		$area = $ni.SelectSingleNode("d:area", $nsMgr)
+		$beginRow = [int]$area.SelectSingleNode("d:beginRow", $nsMgr).InnerText
+		$endRow = [int]$area.SelectSingleNode("d:endRow", $nsMgr).InnerText
+
+		# Check row bounds (skip -1 which means "all")
+		if ($beginRow -ne -1 -and $beginRow -ge $docHeight) {
+			Report-Error "Area '$name': beginRow=$beginRow >= height=$docHeight"
+		}
+		if ($endRow -ne -1 -and $endRow -ge $docHeight) {
+			Report-Error "Area '$name': endRow=$endRow >= height=$docHeight"
+		}
+
+		# Check columnsID reference
+		$colsIdNode = $area.SelectSingleNode("d:columnsID", $nsMgr)
+		if ($colsIdNode) {
+			$colsId = $colsIdNode.InnerText
+			if (-not $columnSets.ContainsKey($colsId)) {
+				Report-Error "Area '$name': columnsID '$($colsId.Substring(0,8))...' not found"
+			}
+		}
+	}
+}
+
+# --- Check 9: merge bounds ---
+
+foreach ($merge in $root.SelectNodes("d:merge", $nsMgr)) {
+	if ($stopped) { break }
+
+	$r = [int]$merge.SelectSingleNode("d:r", $nsMgr).InnerText
+	$c = [int]$merge.SelectSingleNode("d:c", $nsMgr).InnerText
+	$wNode = $merge.SelectSingleNode("d:w", $nsMgr)
+	$hNode = $merge.SelectSingleNode("d:h", $nsMgr)
+
+	# r=-1 means all rows, skip bound check
+	if ($r -ne -1 -and $r -ge $docHeight) {
+		Report-Error "Merge at row=${r}, col=${c}: row >= height ($docHeight)"
+	}
+
+	if ($hNode -and $r -ne -1) {
+		$h = [int]$hNode.InnerText
+		if (($r + $h) -ge $docHeight) {
+			Report-Error "Merge at row=${r}: extends to row $($r + $h) >= height ($docHeight)"
+		}
+	}
+
+	# Check columnsID in merge
+	$colsIdNode = $merge.SelectSingleNode("d:columnsID", $nsMgr)
+	if ($colsIdNode) {
+		$colsId = $colsIdNode.InnerText
+		if (-not $columnSets.ContainsKey($colsId)) {
+			Report-Error "Merge at row=${r}, col=${c}: columnsID '$($colsId.Substring(0,8))...' not found"
+		}
+	}
+}
+
+# --- Check 12: drawing picture indices ---
+
+foreach ($drawing in $root.SelectNodes("d:drawing", $nsMgr)) {
+	if ($stopped) { break }
+
+	$picIdxNode = $drawing.SelectSingleNode("d:pictureIndex", $nsMgr)
+	if ($picIdxNode) {
+		$picIdx = [int]$picIdxNode.InnerText
+		if ($picIdx -gt $pictureCount) {
+			$drawId = $drawing.SelectSingleNode("d:id", $nsMgr).InnerText
+			Report-Error "Drawing id=${drawId}: pictureIndex=$picIdx > picture count ($pictureCount)"
+		}
+	}
+	# Оформление рисунка — такая же запись палитры форматов, как у ячейки, и висячая
+	# ссылка на неё так же валит загрузку макета.
+	$fmtIdxNode = $drawing.SelectSingleNode("d:formatIndex", $nsMgr)
+	if ($fmtIdxNode) {
+		$fmtIdx = [int]$fmtIdxNode.InnerText
+		if ($fmtIdx -gt $formatCount) {
+			$drawId = $drawing.SelectSingleNode("d:id", $nsMgr).InnerText
+			Report-Error "Drawing id=${drawId}: formatIndex=$fmtIdx > format palette size ($formatCount)"
+		}
+	}
+}
+
+# --- Check 13: value cells (input fields) ---
+# Свойства containsValue/valueType/controlType принадлежат ЯЧЕЙКЕ: на корпусе ERP
+# (370 197 ссылок) на такие записи палитры ссылаются только <f>, ни строка, ни колонка,
+# ни defaultFormatIndex. Ячейка со значением текста не несёт — ни одна из 370 197.
+
+$valueFormatIdx = @{}
+$controlGuids = @('381ed624-9217-4e63-85db-c4c3cb87daae', '35af3d93-d7c7-4a2e-a8eb-bac87a1a3f26')
+for ($i = 0; $i -lt $formatNodes.Count; $i++) {
+	$fmt = $formatNodes[$i]
+	$contains = $fmt.SelectSingleNode("d:containsValue", $nsMgr)
+	$vt = $fmt.SelectSingleNode("d:valueType", $nsMgr)
+	if (-not $contains -and -not $vt) { continue }
+	$num = $i + 1
+	$valueFormatIdx[$num] = $true
+	if (-not $contains -or $contains.InnerText.Trim() -cne 'true') {
+		Report-Error "Format ${num}: <valueType> without <containsValue>true</containsValue>"
+	} elseif (-not $vt) {
+		Report-Error "Format ${num}: <containsValue> without <valueType>"
+	}
+	if ($vt) {
+		foreach ($child in $vt.ChildNodes) {
+			if ($child.NodeType -ne [System.Xml.XmlNodeType]::Element) { continue }
+			$tag = $child.get_LocalName()
+			if ($tag -cne 'Type' -and $tag -cne 'TypeSet' -and $tag -notlike '*Qualifiers') {
+				Report-Error "Format ${num}: unexpected <$tag> inside <valueType>"
+			}
+		}
+	}
+	$ctl = $fmt.SelectSingleNode("d:controlType", $nsMgr)
+	if ($ctl -and ($controlGuids -notcontains $ctl.InnerText.Trim().ToLowerInvariant())) {
+		Report-Warn "Format ${num}: unknown controlType $($ctl.InnerText.Trim())"
+	}
+	# Флажок Конфигуратор предлагает только для булева и числа. Движок принимает его на любом
+	# типе (проверено сборкой EPF и обратной выгрузкой), поэтому это предупреждение, а не ошибка:
+	# собрать такую ячейку в Конфигураторе нельзя, и почти наверняка это описка.
+	if ($ctl -and $ctl.InnerText.Trim().ToLowerInvariant() -ceq $controlGuids[1] -and $vt) {
+		$kinds = @()
+		foreach ($child in $vt.ChildNodes) {
+			if ($child.NodeType -ne [System.Xml.XmlNodeType]::Element) { continue }
+			if ($child.get_LocalName() -ceq 'Type') { $kinds += $child.InnerText.Trim() }
+		}
+		if (-not (($kinds.Count -eq 1) -and ($kinds[0] -ceq 'xs:boolean' -or $kinds[0] -ceq 'xs:decimal'))) {
+			Report-Warn "Format ${num}: checkbox control on a type other than Boolean or Number"
+		}
+	}
+}
+
+foreach ($ri in $root.SelectNodes("d:rowsItem", $nsMgr)) {
+	if ($stopped) { break }
+	$row = $ri.SelectSingleNode("d:row", $nsMgr)
+	if (-not $row) { continue }
+	$idxNode = $ri.SelectSingleNode("d:index", $nsMgr)
+	$rn = if ($idxNode) { $idxNode.InnerText } else { '?' }
+	$rowFmt = $row.SelectSingleNode("d:formatIndex", $nsMgr)
+	if ($rowFmt -and $valueFormatIdx.ContainsKey([int]$rowFmt.InnerText)) {
+		Report-Warn "Row ${rn}: formatIndex points to a value format (cell-only property)"
+	}
+	foreach ($cGroup in $row.SelectNodes("d:c", $nsMgr)) {
+		$cell = $cGroup.SelectSingleNode("d:c", $nsMgr)
+		if (-not $cell) { continue }
+		$fNode = $cell.SelectSingleNode("d:f", $nsMgr)
+		if (-not $fNode -or -not $valueFormatIdx.ContainsKey([int]$fNode.InnerText)) {
+			# Значение и настройки элемента управления бывают только у ячейки-поля ввода:
+			# на корпусе ни одного <v> и ни одного <control> в обычной ячейке.
+			if ($cell.SelectSingleNode("d:v", $nsMgr)) {
+				Report-Error "Row ${rn}: cell carries value but its format has no containsValue"
+			}
+			if ($cell.SelectSingleNode("d:control", $nsMgr)) {
+				Report-Error "Row ${rn}: cell carries control settings but its format has no containsValue"
+			}
+			continue
+		}
+		if ($cell.SelectSingleNode("d:tl", $nsMgr)) {
+			Report-Error "Row ${rn}: cell contains a value and text at the same time"
+		}
+	}
+}
+foreach ($cols in $root.SelectNodes("d:columns", $nsMgr)) {
+	foreach ($ci in $cols.SelectNodes("d:columnsItem", $nsMgr)) {
+		$col = $ci.SelectSingleNode("d:column", $nsMgr)
+		if (-not $col) { continue }
+		$fmtNode = $col.SelectSingleNode("d:formatIndex", $nsMgr)
+		if ($fmtNode -and $valueFormatIdx.ContainsKey([int]$fmtNode.InnerText)) {
+			$colIdxNode = $ci.SelectSingleNode("d:index", $nsMgr)
+			$colIdxText = if ($colIdxNode) { $colIdxNode.InnerText } else { '?' }
+			Report-Warn "Column ${colIdxText}: formatIndex points to a value format (cell-only property)"
+		}
+	}
+}
+$dfi = $root.SelectSingleNode("d:defaultFormatIndex", $nsMgr)
+if ($dfi -and $valueFormatIdx.ContainsKey([int]$dfi.InnerText)) {
+	Report-Warn "defaultFormatIndex points to a value format (cell-only property)"
+}
+if ($valueFormatIdx.Count -gt 0) {
+	Report-OK "Value cells: $($valueFormatIdx.Count) value formats"
+}
+
+# --- Check 14: группировки строк и колонок ---
+# Диапазоны группировок либо вложены, либо не пересекаются: на корпусе 40 620 886 пар
+# непересекающихся и 599 958 вложенных, частичных пересечений нет ни одного.
+
+foreach ($tag in @('vg', 'hg')) {
+	$ranges = @()
+	foreach ($g in $root.SelectNodes("d:$tag", $nsMgr)) {
+		$bNode = $g.SelectSingleNode("d:b", $nsMgr)
+		if (-not $bNode) {
+			Report-Error "Group <$tag>: <b> is missing"
+			continue
+		}
+		$b = [int]$bNode.InnerText
+		$eNode = $g.SelectSingleNode("d:e", $nsMgr)
+		$e = if ($eNode) { [int]$eNode.InnerText } else { $b }
+		if ($e -lt $b) { Report-Error "Group <$tag> $b..${e}: range is reversed" }
+		if ($tag -ceq 'vg' -and $docHeight -gt 0 -and $b -ge $docHeight) {
+			Report-Warn "Group <vg> starts at row $b, beyond document height ($docHeight)"
+		}
+		$ranges += ,@($b, $e)
+	}
+	for ($i = 0; $i -lt $ranges.Count; $i++) {
+		for ($j = $i + 1; $j -lt $ranges.Count; $j++) {
+			$a = $ranges[$i]; $c = $ranges[$j]
+			if ($a[1] -lt $c[0] -or $c[1] -lt $a[0]) { continue }
+			if (($a[0] -le $c[0] -and $c[1] -le $a[1]) -or ($c[0] -le $a[0] -and $a[1] -le $c[1])) { continue }
+			Report-Error "Groups <$tag> $($a[0])..$($a[1]) and $($c[0])..$($c[1]) overlap partially"
+		}
+	}
+	if ($ranges.Count -gt 0) { Report-OK "Groups <$tag>: $($ranges.Count)" }
+}
+
+# --- Summary ---
+
+$checks = $script:okCount + $errors + $warnings
+
+if ($errors -eq 0 -and $warnings -eq 0 -and -not $Detailed) {
+	Write-Host "=== Validation OK: Template.$templateName ($checks checks) ==="
+} else {
+	Write-Host ""
+
+	if ($stopped) {
+		Write-Host "Stopped after $MaxErrors errors. Fix and re-run."
+	}
+
+	Write-Host "=== Result: $errors errors, $warnings warnings ($checks checks) ==="
+}
+
+if ($errors -gt 0) {
+	exit 1
+} else {
+	exit 0
+}

.github/skills/mxl-validate/scripts/mxl-validate.py

@@ -0,0 +1,551 @@
+#!/usr/bin/env python3
+# mxl-validate v1.7 — Validate 1C spreadsheet document Template.xml
+# Source: https://github.com/Nikolay-Shirokov/cc-1c-skills
+"""Validates spreadsheet Template.xml: height, palette refs, column/row indices, areas, merges."""
+import sys, os, argparse
+from lxml import etree
+
+# Регистронезависимый ввод — паритет с PS1: в PowerShell имена параметров и [ValidateSet]
+# регистр не различают, в argparse совпадение точное.
+def ci_parse_args(parser, argv=None):
+    """parse_args по правилам PS: имена параметров и значения choices регистронезависимы."""
+    argv = list(sys.argv[1:] if argv is None else argv)
+    names = {s.lower(): s for a in parser._actions for s in a.option_strings}
+    for i, tok in enumerate(argv):
+        if tok.startswith('-') and tok.lower() in names:
+            argv[i] = names[tok.lower()]
+    # choices — зеркало [ValidateSet]; канонизируем ДО разбора, иначе argparse отвергнет регистр
+    choice_map = {}
+    for a in parser._actions:
+        if a.choices:
+            for s in a.option_strings:
+                choice_map[s] = {str(c).lower(): c for c in a.choices}
+    for i in range(len(argv) - 1):
+        m = choice_map.get(argv[i])
+        if m and argv[i + 1].lower() in m:
+            argv[i + 1] = m[argv[i + 1].lower()]
+    return parser.parse_args(argv)
+
+
+NS_D   = 'http://v8.1c.ru/8.2/data/spreadsheet'
+NS_V8  = 'http://v8.1c.ru/8.1/data/core'
+NS_XSI = 'http://www.w3.org/2001/XMLSchema-instance'
+
+NS = {
+    'd':   NS_D,
+    'v8':  NS_V8,
+    'xsi': NS_XSI,
+}
+
+
+class Reporter:
+    def __init__(self, max_errors, detailed=False):
+        self.errors = 0
+        self.warnings = 0
+        self.ok_count = 0
+        self.stopped = False
+        self.max_errors = max_errors
+        self.detailed = detailed
+        self.lines = []
+
+    def ok(self, msg):
+        self.ok_count += 1
+        if self.detailed:
+            self.lines.append(f'[OK]    {msg}')
+
+    def error(self, msg):
+        self.errors += 1
+        self.lines.append(f'[ERROR] {msg}')
+        if self.errors >= self.max_errors:
+            self.stopped = True
+
+    def warn(self, msg):
+        self.warnings += 1
+        self.lines.append(f'[WARN]  {msg}')
+
+
+def int_text(node):
+    """Return int from node text, or 0 if None."""
+    if node is not None and node.text:
+        return int(node.text)
+    return 0
+
+
+def main():
+    sys.stdout.reconfigure(encoding="utf-8")
+    sys.stderr.reconfigure(encoding="utf-8")
+    parser = argparse.ArgumentParser(
+        description='Validate 1C spreadsheet document Template.xml', allow_abbrev=False
+    )
+    parser.add_argument('-TemplatePath', '-Path', dest='TemplatePath', default='')
+    parser.add_argument('-ProcessorName', dest='ProcessorName', default='')
+    parser.add_argument('-TemplateName', dest='TemplateName', default='')
+    parser.add_argument('-SrcDir', dest='SrcDir', default='src')
+    parser.add_argument('-Detailed', action='store_true')
+    parser.add_argument('-MaxErrors', dest='MaxErrors', type=int, default=20)
+    args = ci_parse_args(parser)
+
+    template_path = args.TemplatePath
+    processor_name = args.ProcessorName
+    template_name_arg = args.TemplateName
+    src_dir = args.SrcDir
+    detailed = args.Detailed
+    max_errors = args.MaxErrors
+
+    # --- Resolve template path ---
+    if not template_path:
+        if not processor_name or not template_name_arg:
+            print('Specify -TemplatePath or both -ProcessorName and -TemplateName', file=sys.stderr)
+            sys.exit(1)
+        template_path = os.path.join(src_dir, processor_name, 'Templates',
+                                     template_name_arg, 'Ext', 'Template.xml')
+
+    if not os.path.isabs(template_path):
+        template_path = os.path.join(os.getcwd(), template_path)
+
+    # A: Directory → Ext/Template.xml
+    if os.path.isdir(template_path):
+        template_path = os.path.join(template_path, 'Ext', 'Template.xml')
+    # B1: Missing Ext/ (e.g. Templates/Макет/Template.xml → Templates/Макет/Ext/Template.xml)
+    if not os.path.exists(template_path):
+        fn = os.path.basename(template_path)
+        if fn == 'Template.xml':
+            c = os.path.join(os.path.dirname(template_path), 'Ext', fn)
+            if os.path.exists(c):
+                template_path = c
+    # B2: Descriptor (Templates/Макет.xml → Templates/Макет/Ext/Template.xml)
+    if not os.path.exists(template_path) and template_path.endswith('.xml'):
+        stem = os.path.splitext(os.path.basename(template_path))[0]
+        parent = os.path.dirname(template_path)
+        c = os.path.join(parent, stem, 'Ext', 'Template.xml')
+        if os.path.exists(c):
+            template_path = c
+
+    if not os.path.exists(template_path):
+        print(f'File not found: {template_path}', file=sys.stderr)
+        sys.exit(1)
+
+    resolved_path = os.path.abspath(template_path)
+
+    # --- Load XML ---
+    xml_parser = etree.XMLParser(remove_blank_text=False)
+    xml_doc = etree.parse(resolved_path, xml_parser)
+    root = xml_doc.getroot()
+
+    r = Reporter(max_errors, detailed)
+
+    # Derive template name from path: .../Templates/<Name>/Ext/Template.xml
+    # Go up 2 levels from Template.xml -> Ext -> <Name>
+    template_display_name = os.path.basename(os.path.dirname(os.path.dirname(resolved_path)))
+    r.lines.append(f'=== Validation: Template.{template_display_name} ===')
+    r.lines.append('')
+
+    # --- Collect palettes ---
+    line_nodes = root.findall(f'{{{NS_D}}}line')
+    line_count = len(line_nodes)
+
+    font_nodes = [node for node in root if isinstance(node.tag, str) and etree.QName(node.tag).localname == 'font']
+    font_count = len(font_nodes)
+
+    format_nodes = [node for node in root if isinstance(node.tag, str) and etree.QName(node.tag).localname == 'format']
+    format_count = len(format_nodes)
+
+    picture_nodes = root.findall(f'{{{NS_D}}}picture')
+    picture_count = len(picture_nodes)
+
+    # --- Collect column sets ---
+    column_sets = {}  # id -> size
+    default_col_count = 0
+
+    for cols in root.findall(f'{{{NS_D}}}columns'):
+        size_node = cols.find(f'{{{NS_D}}}size')
+        id_node = cols.find(f'{{{NS_D}}}id')
+        size = int_text(size_node)
+
+        if id_node is not None and id_node.text:
+            column_sets[id_node.text] = size
+        else:
+            default_col_count = size
+
+    # --- Check 1: height vs actual rows ---
+    row_nodes = root.findall(f'{{{NS_D}}}rowsItem')
+    height_node = root.find(f'{{{NS_D}}}height')
+    doc_height = int_text(height_node)
+
+    max_row_index = -1
+    for ri in row_nodes:
+        idx_node = ri.find(f'{{{NS_D}}}index')
+        if idx_node is not None and idx_node.text:
+            idx = int(idx_node.text)
+            if idx > max_row_index:
+                max_row_index = idx
+
+    expected_min_height = max_row_index + 1
+    if doc_height >= expected_min_height:
+        r.ok(f'height ({doc_height}) >= max row index + 1 ({expected_min_height}), rowsItem count={len(row_nodes)}')
+    else:
+        r.error(f'height={doc_height} but max row index={max_row_index} (need at least {expected_min_height})')
+
+    # --- Check 2: vgRows <= height ---
+    vg_rows_node = root.find(f'{{{NS_D}}}vgRows')
+    if vg_rows_node is not None:
+        vg_rows = int_text(vg_rows_node)
+        if vg_rows <= doc_height:
+            r.ok(f'vgRows ({vg_rows}) <= height ({doc_height})')
+        else:
+            r.warn(f'vgRows ({vg_rows}) > height ({doc_height})')
+
+    # --- Build row data for checks ---
+    max_format_ref = 0
+    max_font_ref = 0
+    max_line_ref = 0
+
+    # Check format palette references in formats (font, border indices)
+    for fmt in format_nodes:
+        font_idx_node = fmt.find(f'{{{NS_D}}}font')
+        if font_idx_node is not None and font_idx_node.text:
+            val = int(font_idx_node.text)
+            if val > max_font_ref:
+                max_font_ref = val
+
+        for border_name in ('leftBorder', 'topBorder', 'rightBorder', 'bottomBorder', 'drawingBorder'):
+            border_node = fmt.find(f'{{{NS_D}}}{border_name}')
+            if border_node is not None and border_node.text:
+                val = int(border_node.text)
+                if val > max_line_ref:
+                    max_line_ref = val
+
+    # --- Check 10: font indices in formats ---
+    if font_count > 0:
+        if max_font_ref < font_count:
+            r.ok(f'Font refs: max={max_font_ref}, palette size={font_count}')
+        else:
+            r.error(f'Font index {max_font_ref} exceeds palette size ({font_count})')
+    elif max_font_ref > 0:
+        r.error(f'Font index {max_font_ref} referenced but no fonts defined')
+    else:
+        # Макет без единого шрифта — норма (ячейки-поля ввода, макеты без оформления).
+        # Строку печатаем как ps1-порт: раньше py молчал, и на таком макете число проверок
+        # у портов расходилось.
+        r.ok('No font references')
+
+    # --- Check 11: line/border indices in formats ---
+    if line_count > 0:
+        if max_line_ref < line_count:
+            r.ok(f'Line/border refs: max={max_line_ref}, palette size={line_count}')
+        else:
+            r.error(f'Line index {max_line_ref} exceeds palette size ({line_count})')
+    elif max_line_ref > 0:
+        r.error(f'Line index {max_line_ref} referenced but no lines defined')
+    else:
+        r.ok('No line/border references')
+
+    # --- Check 3, 4, 5, 6: row/cell checks ---
+    max_cell_format_ref = 0
+    max_row_format_ref = 0
+    max_default_col_idx = 0
+    row_index = 0
+
+    for ri in row_nodes:
+        if r.stopped:
+            break
+
+        idx_node = ri.find(f'{{{NS_D}}}index')
+        if idx_node is not None and idx_node.text:
+            row_index = int(idx_node.text)
+
+        row = ri.find(f'{{{NS_D}}}row')
+        if row is None:
+            row_index += 1
+            continue
+
+        # Row formatIndex
+        row_fmt_node = row.find(f'{{{NS_D}}}formatIndex')
+        if row_fmt_node is not None and row_fmt_node.text:
+            val = int(row_fmt_node.text)
+            if val > max_row_format_ref:
+                max_row_format_ref = val
+            if val > format_count:
+                r.error(f'Row {row_index}: formatIndex={val} > format palette size ({format_count})')
+
+        # Check columnsID
+        row_cols_id = None
+        cols_id_node = row.find(f'{{{NS_D}}}columnsID')
+        if cols_id_node is not None and cols_id_node.text:
+            row_cols_id = cols_id_node.text
+            if row_cols_id not in column_sets:
+                r.error(f"Row {row_index}: columnsID '{row_cols_id[:8]}...' not found in column sets")
+
+        # Determine column count for this row
+        row_col_count = default_col_count
+        if row_cols_id and row_cols_id in column_sets:
+            row_col_count = column_sets[row_cols_id]
+
+        # Cell checks
+        for c_group in row.findall(f'{{{NS_D}}}c'):
+            i_node = c_group.find(f'{{{NS_D}}}i')
+            col_idx = None
+            if i_node is not None and i_node.text:
+                col_idx = int(i_node.text)
+                # Track max index for default column set only
+                if row_cols_id is None and col_idx > max_default_col_idx:
+                    max_default_col_idx = col_idx
+                # Check against row's column count
+                if row_col_count > 0 and col_idx >= row_col_count:
+                    r.error(f'Row {row_index}: column index {col_idx} >= column count ({row_col_count})')
+
+            cell = c_group.find(f'{{{NS_D}}}c')
+            if cell is not None:
+                f_node = cell.find(f'{{{NS_D}}}f')
+                if f_node is not None and f_node.text:
+                    val = int(f_node.text)
+                    if val > max_cell_format_ref:
+                        max_cell_format_ref = val
+                    if val > format_count:
+                        r.error(f'Row {row_index}: cell format index {val} > format palette size ({format_count})')
+                # Примечание — четвёртый владелец формата, и его ссылка тоже бывает битой.
+                note = cell.find(f'{{{NS_D}}}note')
+                if note is not None:
+                    nf = note.find(f'{{{NS_D}}}formatIndex')
+                    if nf is not None and nf.text and int(nf.text) > format_count:
+                        r.error(f'Row {row_index}: note format index {nf.text}'
+                                f' > format palette size ({format_count})')
+
+        row_index += 1
+
+    # Summary checks for format refs
+    if not r.stopped:
+        if max_cell_format_ref <= format_count and max_row_format_ref <= format_count:
+            r.ok(f'Format refs: max cell={max_cell_format_ref}, max row={max_row_format_ref}, palette size={format_count}')
+
+    # Check column format indices
+    for cols in root.findall(f'{{{NS_D}}}columns'):
+        if r.stopped:
+            break
+        for ci in cols.findall(f'{{{NS_D}}}columnsItem'):
+            col = ci.find(f'{{{NS_D}}}column')
+            if col is not None:
+                fmt_node = col.find(f'{{{NS_D}}}formatIndex')
+                if fmt_node is not None and fmt_node.text:
+                    val = int(fmt_node.text)
+                    if val > format_count:
+                        col_idx_node = ci.find(f'{{{NS_D}}}index')
+                        col_idx_text = col_idx_node.text if col_idx_node is not None else '?'
+                        r.error(f'Column {col_idx_text}: formatIndex={val} > format palette size ({format_count})')
+
+    # --- Check 5: column index summary ---
+    if not r.stopped:
+        r.ok(f'Column indices: max in default set={max_default_col_idx}, default column count={default_col_count}')
+
+    # --- Check 7, 8: named areas ---
+    for ni in root.findall(f'{{{NS_D}}}namedItem'):
+        if r.stopped:
+            break
+
+        ni_type = ni.get(f'{{{NS_XSI}}}type', '')
+        name_node = ni.find(f'{{{NS_D}}}name')
+        name = name_node.text if name_node is not None else ''
+
+        if 'NamedItemCells' in ni_type:
+            area = ni.find(f'{{{NS_D}}}area')
+            if area is None:
+                continue
+            begin_row = int_text(area.find(f'{{{NS_D}}}beginRow'))
+            end_row = int_text(area.find(f'{{{NS_D}}}endRow'))
+
+            # Check row bounds (skip -1 which means "all")
+            if begin_row != -1 and begin_row >= doc_height:
+                r.error(f"Area '{name}': beginRow={begin_row} >= height={doc_height}")
+            if end_row != -1 and end_row >= doc_height:
+                r.error(f"Area '{name}': endRow={end_row} >= height={doc_height}")
+
+            # Check columnsID reference
+            cols_id_node = area.find(f'{{{NS_D}}}columnsID')
+            if cols_id_node is not None and cols_id_node.text:
+                cols_id = cols_id_node.text
+                if cols_id not in column_sets:
+                    r.error(f"Area '{name}': columnsID '{cols_id[:8]}...' not found")
+
+    # --- Check 9: merge bounds ---
+    for merge in root.findall(f'{{{NS_D}}}merge'):
+        if r.stopped:
+            break
+
+        merge_r = int_text(merge.find(f'{{{NS_D}}}r'))
+        merge_c = int_text(merge.find(f'{{{NS_D}}}c'))
+        w_node = merge.find(f'{{{NS_D}}}w')
+        h_node = merge.find(f'{{{NS_D}}}h')
+
+        # r=-1 means all rows, skip bound check
+        if merge_r != -1 and merge_r >= doc_height:
+            r.error(f'Merge at row={merge_r}, col={merge_c}: row >= height ({doc_height})')
+
+        if h_node is not None and merge_r != -1:
+            h = int_text(h_node)
+            if (merge_r + h) >= doc_height:
+                r.error(f'Merge at row={merge_r}: extends to row {merge_r + h} >= height ({doc_height})')
+
+        # Check columnsID in merge
+        cols_id_node = merge.find(f'{{{NS_D}}}columnsID')
+        if cols_id_node is not None and cols_id_node.text:
+            cols_id = cols_id_node.text
+            if cols_id not in column_sets:
+                r.error(f"Merge at row={merge_r}, col={merge_c}: columnsID '{cols_id[:8]}...' not found")
+
+    # --- Check 12: drawing picture indices ---
+    for drawing in root.findall(f'{{{NS_D}}}drawing'):
+        if r.stopped:
+            break
+
+        pic_idx_node = drawing.find(f'{{{NS_D}}}pictureIndex')
+        if pic_idx_node is not None and pic_idx_node.text:
+            pic_idx = int(pic_idx_node.text)
+            if pic_idx > picture_count:
+                draw_id_node = drawing.find(f'{{{NS_D}}}id')
+                draw_id = draw_id_node.text if draw_id_node is not None else '?'
+                r.error(f'Drawing id={draw_id}: pictureIndex={pic_idx} > picture count ({picture_count})')
+
+        # Оформление рисунка — такая же запись палитры форматов, как у ячейки, и висячая
+        # ссылка на неё так же валит загрузку макета.
+        fmt_idx_node = drawing.find(f'{{{NS_D}}}formatIndex')
+        if fmt_idx_node is not None and fmt_idx_node.text:
+            fmt_idx = int(fmt_idx_node.text)
+            if fmt_idx > format_count:
+                draw_id_node = drawing.find(f'{{{NS_D}}}id')
+                draw_id = draw_id_node.text if draw_id_node is not None else '?'
+                r.error(f'Drawing id={draw_id}: formatIndex={fmt_idx} > format palette size ({format_count})')
+
+    # --- Check 13: value cells (input fields) ---
+    # Свойства containsValue/valueType/controlType принадлежат ЯЧЕЙКЕ: на корпусе ERP
+    # (370 197 ссылок) на такие записи палитры ссылаются только <f>, ни строка, ни колонка,
+    # ни defaultFormatIndex. Ячейка со значением текста не несёт — ни одна из 370 197.
+    value_format_idx = set()
+    control_guids = ('381ed624-9217-4e63-85db-c4c3cb87daae', '35af3d93-d7c7-4a2e-a8eb-bac87a1a3f26')
+    for i, fmt in enumerate(format_nodes, start=1):
+        contains = fmt.find(f'{{{NS_D}}}containsValue')
+        vt = fmt.find(f'{{{NS_D}}}valueType')
+        if contains is None and vt is None:
+            continue
+        value_format_idx.add(i)
+        if contains is None or (contains.text or '').strip() != 'true':
+            r.error(f'Format {i}: <valueType> without <containsValue>true</containsValue>')
+        elif vt is None:
+            r.error(f'Format {i}: <containsValue> without <valueType>')
+        if vt is not None:
+            for child in vt:
+                tag = child.tag.split('}')[-1]
+                if tag not in ('Type', 'TypeSet') and not tag.endswith('Qualifiers'):
+                    r.error(f'Format {i}: unexpected <{tag}> inside <valueType>')
+        ctl = fmt.find(f'{{{NS_D}}}controlType')
+        if ctl is not None and (ctl.text or '').strip().lower() not in control_guids:
+            r.warn(f'Format {i}: unknown controlType {(ctl.text or "").strip()}')
+        # Флажок Конфигуратор предлагает только для булева и числа. Движок принимает его на
+        # любом типе (проверено сборкой EPF и обратной выгрузкой), поэтому это предупреждение,
+        # а не ошибка: собрать такую ячейку в Конфигураторе нельзя, и почти наверняка это описка.
+        if ctl is not None and (ctl.text or '').strip().lower() == control_guids[1] and vt is not None:
+            kinds = [(t.text or '').strip() for t in vt if t.tag.endswith('}Type')]
+            if kinds != ['xs:boolean'] and kinds != ['xs:decimal']:
+                r.warn(f'Format {i}: checkbox control on a type other than Boolean or Number')
+
+    for ri in root.findall(f'{{{NS_D}}}rowsItem'):
+        if r.stopped:
+            break
+        row = ri.find(f'{{{NS_D}}}row')
+        if row is None:
+            continue
+        idx_node = ri.find(f'{{{NS_D}}}index')
+        rn = idx_node.text if idx_node is not None else '?'
+        row_fmt = row.find(f'{{{NS_D}}}formatIndex')
+        if row_fmt is not None and row_fmt.text and int(row_fmt.text) in value_format_idx:
+            r.warn(f'Row {rn}: formatIndex points to a value format (cell-only property)')
+        for c_group in row.findall(f'{{{NS_D}}}c'):
+            cell = c_group.find(f'{{{NS_D}}}c')
+            if cell is None:
+                continue
+            f_node = cell.find(f'{{{NS_D}}}f')
+            if f_node is None or not f_node.text or int(f_node.text) not in value_format_idx:
+                continue
+            if cell.find(f'{{{NS_D}}}tl') is not None:
+                r.error(f'Row {rn}: cell contains a value and text at the same time')
+    # Значение и настройки элемента управления бывают только у ячейки-поля ввода: на корпусе
+    # ни одного <v> и ни одного <control> в обычной ячейке.
+    for ri in root.findall(f'{{{NS_D}}}rowsItem'):
+        if r.stopped:
+            break
+        row = ri.find(f'{{{NS_D}}}row')
+        if row is None:
+            continue
+        idx_node = ri.find(f'{{{NS_D}}}index')
+        rn = idx_node.text if idx_node is not None else '?'
+        for c_group in row.findall(f'{{{NS_D}}}c'):
+            cell = c_group.find(f'{{{NS_D}}}c')
+            if cell is None:
+                continue
+            f_node = cell.find(f'{{{NS_D}}}f')
+            is_value = f_node is not None and f_node.text and int(f_node.text) in value_format_idx
+            if is_value:
+                continue
+            for tag, what in (('v', 'value'), ('control', 'control settings')):
+                if cell.find(f'{{{NS_D}}}{tag}') is not None:
+                    r.error(f'Row {rn}: cell carries {what} but its format has no containsValue')
+    for cols in root.findall(f'{{{NS_D}}}columns'):
+        for ci in cols.findall(f'{{{NS_D}}}columnsItem'):
+            col = ci.find(f'{{{NS_D}}}column')
+            if col is None:
+                continue
+            fmt_node = col.find(f'{{{NS_D}}}formatIndex')
+            if fmt_node is not None and fmt_node.text and int(fmt_node.text) in value_format_idx:
+                col_idx_node = ci.find(f'{{{NS_D}}}index')
+                col_idx_text = col_idx_node.text if col_idx_node is not None else '?'
+                r.warn(f'Column {col_idx_text}: formatIndex points to a value format (cell-only property)')
+    dfi = root.find(f'{{{NS_D}}}defaultFormatIndex')
+    if dfi is not None and dfi.text and int(dfi.text) in value_format_idx:
+        r.warn('defaultFormatIndex points to a value format (cell-only property)')
+    if value_format_idx:
+        r.ok(f'Value cells: {len(value_format_idx)} value formats')
+
+    # --- Check 14: группировки строк и колонок ---
+    # Диапазоны группировок либо вложены, либо не пересекаются: на корпусе 40 620 886 пар
+    # непересекающихся и 599 958 вложенных, частичных пересечений нет ни одного.
+    for tag in ('vg', 'hg'):
+        ranges = []
+        for g in root.findall(f'{{{NS_D}}}{tag}'):
+            b_node = g.find(f'{{{NS_D}}}b')
+            if b_node is None or not b_node.text:
+                r.error(f'Group <{tag}>: <b> is missing')
+                continue
+            b = int(b_node.text)
+            e_node = g.find(f'{{{NS_D}}}e')
+            e = int(e_node.text) if e_node is not None and e_node.text else b
+            if e < b:
+                r.error(f'Group <{tag}> {b}..{e}: range is reversed')
+            if tag == 'vg' and doc_height > 0 and b >= doc_height:
+                r.warn(f'Group <vg> starts at row {b}, beyond document height ({doc_height})')
+            ranges.append((b, e))
+        for i in range(len(ranges)):
+            for j in range(i + 1, len(ranges)):
+                a, c = ranges[i], ranges[j]
+                if a[1] < c[0] or c[1] < a[0]:
+                    continue
+                if (a[0] <= c[0] and c[1] <= a[1]) or (c[0] <= a[0] and a[1] <= c[1]):
+                    continue
+                r.error(f'Groups <{tag}> {a[0]}..{a[1]} and {c[0]}..{c[1]} overlap partially')
+        if ranges:
+            r.ok(f'Groups <{tag}>: {len(ranges)}')
+
+    # --- Finalize ---
+    checks = r.ok_count + r.errors + r.warnings
+    if r.errors == 0 and r.warnings == 0 and not detailed:
+        result = f'=== Validation OK: Template.{template_display_name} ({checks} checks) ==='
+    else:
+        r.lines.append('')
+        r.lines.append(f'=== Result: {r.errors} errors, {r.warnings} warnings ({checks} checks) ===')
+        result = '\n'.join(r.lines)
+
+    print(result)
+
+    sys.exit(1 if r.errors > 0 else 0)
+
+
+if __name__ == '__main__':
+    main()