Add proofread_markdown skill and Dart wrapper script

.agents/skills/proofread_markdown/SKILL.md

@@ -0,0 +1,53 @@
+---
+name: proofread-markdown
+description: Proofreads Markdown files against Google guidelines.
+---
+
+# Proofreading Markdown
+
+## Overview
+
+Help us transform technical text into clear, concise,
+and machine-readable Markdown. We follow the
+Google Developer Documentation Style Guide.
+
+## Workflow
+
+### 1. Check line length
+
+- [ ] Wrap lines semantically and keep them generally under 80 characters using
+      [scripts/wrap_lines.dart](scripts/wrap_lines.dart)
+- [ ] Verify that the file was wrapped correctly, preserving headings and
+      code blocks
+
+### 2. Fix voice and tone
+
+- [ ] Use active voice.
+- [ ] Use present tense.
+- [ ] Use second person ("you"). No "we".
+
+### 3. Check word choice
+
+- [ ] Replace forbidden terms: "e.g.", "i.e.", "etc.", "should", "would",
+      "could".
+- [ ] Use Oxford comma and American spelling.
+
+### 4. Check style of headings
+
+- [ ] Use Sentence case for headings.
+      - If you change a heading, add an id after the heading like this:
+        "## Configure the project {: #project-configuration }"
+- [ ] Put a blank line after all headings.
+
+### 5. Check style of lists
+
+- [ ] Numbered lists MUST use `1.` for every item to allow easy reordering.
+- [ ] How-to sections must have active voice in headings.
+      - No: "## Project Configuration"
+      - Yes: "## Configure the project"
+
+### 6. Use less words in sections with steps
+
+- [ ] Ask the user if they would like the steps to be more concise. 
+      If yes, update the steps to be brief. Keep it short, simple, and
+      accurate.

.agents/skills/proofread_markdown/scripts/wrap_lines.dart

@@ -0,0 +1,128 @@
+import 'dart:io';
+
+/// A script to wrap markdown lines at 80 columns, preserving headings and code blocks.
+void main(List<String> arguments) {
+  if (arguments.isEmpty) {
+    print('Usage: dart wrap_lines.dart <file_path>');
+    exit(1);
+  }
+
+  final filePath = arguments[0];
+  final file = File(filePath);
+
+  if (!file.existsSync()) {
+    print('File not found: $filePath');
+    exit(1);
+  }
+
+  final content = file.readAsStringSync();
+  final wrappedContent = wrapMarkdown(content);
+  file.writeAsStringSync(wrappedContent);
+
+  print('Successfully wrapped markdown for $filePath');
+}
+
+String wrapMarkdown(String content, {int width = 80}) {
+  final lines = content.split('\n');
+  final output = <String>[];
+  bool inCodeBlock = false;
+
+  for (final line in lines) {
+    final stripped = line.trim();
+
+    // Handle code blocks
+    if (stripped.startsWith('```') || stripped.startsWith('~~~')) {
+      inCodeBlock = !inCodeBlock;
+      output.add(line);
+      continue;
+    }
+
+    if (inCodeBlock) {
+      output.add(line);
+      continue;
+    }
+
+    // Handle headings
+    if (stripped.startsWith('#')) {
+      output.add(line);
+      continue;
+    }
+
+    // Handle empty lines
+    if (stripped.isEmpty) {
+      output.add('');
+      continue;
+    }
+
+    // Handle lists or quotes (matching python script's simple approach)
+    if (stripped.startsWith('-') ||
+        stripped.startsWith('*') ||
+        RegExp(r'^\d+\.').hasMatch(stripped)) {
+      output.addAll(wrapText(line, width));
+      continue;
+    }
+
+    // Normal text paragraph
+    output.addAll(wrapText(line, width));
+  }
+
+  return output.join('\n');
+}
+
+List<String> wrapText(String text, int width) {
+  final leadingSpace = RegExp(r'^\s*').stringMatch(text) ?? '';
+  final content = text.trim();
+
+  if (content.isEmpty) return [leadingSpace];
+
+  // Split by spaces that follow punctuation (.,;:!?)
+  // We use a lookbehind to keep the punctuation with the preceding chunk.
+  final chunks = content.split(RegExp(r'(?<=[.,;:!?])\s+'));
+  final lines = <String>[];
+  var currentLine = leadingSpace;
+
+  for (final chunk in chunks) {
+    // If the chunk itself is longer than width, fallback to word wrapping for this chunk
+    if (chunk.length > width) {
+      // Flush current line if not empty
+      if (currentLine != leadingSpace) {
+        lines.add(currentLine);
+        currentLine = leadingSpace;
+      }
+
+      // Word wrap the chunk
+      final words = chunk.split(RegExp(r'\s+'));
+      for (final word in words) {
+        final space = currentLine == leadingSpace ? '' : ' ';
+        if ((currentLine.length + space.length + word.length) > width) {
+          if (currentLine != leadingSpace) {
+            lines.add(currentLine);
+            currentLine = leadingSpace + word;
+          } else {
+            lines.add(leadingSpace + word);
+            currentLine = leadingSpace;
+          }
+        } else {
+          currentLine += space + word;
+        }
+      }
+      continue;
+    }
+
+    final space = currentLine == leadingSpace ? '' : ' ';
+    if ((currentLine.length + space.length + chunk.length) > width) {
+      if (currentLine != leadingSpace) {
+        lines.add(currentLine);
+      }
+      currentLine = leadingSpace + chunk;
+    } else {
+      currentLine += space + chunk;
+    }
+  }
+
+  if (currentLine != leadingSpace) {
+    lines.add(currentLine);
+  }
+
+  return lines;
+}