Release v1.8.0: Add transcript-fixer skill

## New Skill: transcript-fixer v1.0.0

Correct speech-to-text (ASR/STT) transcription errors through dictionary-based rules and AI-powered corrections with automatic pattern learning.

**Features:**
- Two-stage correction pipeline (dictionary + AI)
- Automatic pattern detection and learning
- Domain-specific dictionaries (general, embodied_ai, finance, medical)
- SQLite-based correction repository
- Team collaboration with import/export
- GLM API integration for AI corrections
- Cost optimization through dictionary promotion

**Use cases:**
- Correcting meeting notes, lecture recordings, or interview transcripts
- Fixing Chinese/English homophone errors and technical terminology
- Building domain-specific correction dictionaries
- Improving transcript accuracy through iterative learning

**Documentation:**
- Complete workflow guides in references/
- SQL query templates
- Troubleshooting guide
- Team collaboration patterns
- API setup instructions

**Marketplace updates:**
- Updated marketplace to v1.8.0
- Added transcript-fixer plugin (category: productivity)
- Updated README.md with skill description and use cases
- Updated CLAUDE.md with skill listing and counts

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

transcript-fixer/scripts/utils/diff_formats/__init__.py

@@ -0,0 +1,18 @@
+"""
+Diff format generators for transcript comparison
+"""
+
+from .unified_format import generate_unified_diff
+from .html_format import generate_html_diff
+from .inline_format import generate_inline_diff
+from .markdown_format import generate_markdown_report
+from .change_extractor import extract_changes, generate_change_summary
+
+__all__ = [
+    'generate_unified_diff',
+    'generate_html_diff',
+    'generate_inline_diff',
+    'generate_markdown_report',
+    'extract_changes',
+    'generate_change_summary',
+]

transcript-fixer/scripts/utils/diff_formats/change_extractor.py

@@ -0,0 +1,102 @@
+#!/usr/bin/env python3
+"""
+Change extraction and summarization
+
+SINGLE RESPONSIBILITY: Extract and summarize changes between text versions
+"""
+
+from __future__ import annotations
+
+import difflib
+
+from .text_splitter import split_into_words
+
+
+def extract_changes(original: str, fixed: str) -> list[dict]:
+    """
+    Extract all changes and return change list
+
+    Args:
+        original: Original text
+        fixed: Fixed text
+
+    Returns:
+        List of change dictionaries with type, context, and content
+    """
+    original_words = split_into_words(original)
+    fixed_words = split_into_words(fixed)
+
+    diff = difflib.SequenceMatcher(None, original_words, fixed_words)
+    changes = []
+
+    for tag, i1, i2, j1, j2 in diff.get_opcodes():
+        if tag == 'replace':
+            original_text = ''.join(original_words[i1:i2])
+            fixed_text = ''.join(fixed_words[j1:j2])
+            changes.append({
+                'type': 'replace',
+                'original': original_text,
+                'fixed': fixed_text,
+                'context_before': ''.join(original_words[max(0, i1-5):i1]),
+                'context_after': ''.join(original_words[i2:min(len(original_words), i2+5)])
+            })
+        elif tag == 'delete':
+            original_text = ''.join(original_words[i1:i2])
+            changes.append({
+                'type': 'delete',
+                'original': original_text,
+                'fixed': '',
+                'context_before': ''.join(original_words[max(0, i1-5):i1]),
+                'context_after': ''.join(original_words[i2:min(len(original_words), i2+5)])
+            })
+        elif tag == 'insert':
+            fixed_text = ''.join(fixed_words[j1:j2])
+            changes.append({
+                'type': 'insert',
+                'original': '',
+                'fixed': fixed_text,
+                'context_before': ''.join(fixed_words[max(0, j1-5):j1]) if j1 > 0 else '',
+                'context_after': ''.join(fixed_words[j2:min(len(fixed_words), j2+5)])
+            })
+
+    return changes
+
+
+def generate_change_summary(changes: list[dict]) -> str:
+    """
+    Generate change summary
+
+    Args:
+        changes: List of change dictionaries
+
+    Returns:
+        Formatted summary string
+    """
+    result = []
+    result.append("=" * 80)
+    result.append(f"修改摘要 (共 {len(changes)} 处修改)")
+    result.append("=" * 80)
+    result.append("")
+
+    for i, change in enumerate(changes, 1):
+        change_type = {
+            'replace': '替换',
+            'delete': '删除',
+            'insert': '添加'
+        }[change['type']]
+
+        result.append(f"[{i}] {change_type}")
+
+        if change['original']:
+            result.append(f"  原文: {change['original']}")
+        if change['fixed']:
+            result.append(f"  修复: {change['fixed']}")
+
+        # Show context
+        context = change['context_before'] + "【修改处】" + change['context_after']
+        if context.strip():
+            result.append(f"  上下文: ...{context}...")
+
+        result.append("")
+
+    return '\n'.join(result)

transcript-fixer/scripts/utils/diff_formats/html_format.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python3
+"""
+HTML diff format generator
+
+SINGLE RESPONSIBILITY: Generate HTML side-by-side comparison
+"""
+
+from __future__ import annotations
+
+import difflib
+
+
+def generate_html_diff(original: str, fixed: str) -> str:
+    """
+    Generate HTML format comparison report (side-by-side)
+
+    Args:
+        original: Original text
+        fixed: Fixed text
+
+    Returns:
+        HTML format string with side-by-side comparison
+    """
+    original_lines = original.splitlines(keepends=True)
+    fixed_lines = fixed.splitlines(keepends=True)
+
+    differ = difflib.HtmlDiff(wrapcolumn=80)
+    html = differ.make_file(
+        original_lines,
+        fixed_lines,
+        fromdesc='原始版本',
+        todesc='修复版本',
+        context=True,
+        numlines=3
+    )
+
+    return html

transcript-fixer/scripts/utils/diff_formats/inline_format.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+"""
+Inline diff format generator
+
+SINGLE RESPONSIBILITY: Generate inline diff with change markers
+"""
+
+from __future__ import annotations
+
+import difflib
+
+from .text_splitter import split_into_words
+
+
+def generate_inline_diff(original: str, fixed: str) -> str:
+    """
+    Generate inline diff marking deletions and additions
+
+    Format:
+        - Normal words: unchanged
+        - Deletions: [-word-]
+        - Additions: [+word+]
+
+    Args:
+        original: Original text
+        fixed: Fixed text
+
+    Returns:
+        Inline diff string with markers
+    """
+    original_words = split_into_words(original)
+    fixed_words = split_into_words(fixed)
+
+    diff = difflib.ndiff(original_words, fixed_words)
+
+    result = []
+    result.append("=" * 80)
+    result.append("行内词语级别对比 (- 删除, + 添加, ? 修改标记)")
+    result.append("=" * 80)
+    result.append("")
+
+    current_line = []
+    for item in diff:
+        marker = item[0]
+        word = item[2:]
+
+        if marker == ' ':
+            current_line.append(word)
+        elif marker == '-':
+            current_line.append(f"[-{word}-]")
+        elif marker == '+':
+            current_line.append(f"[+{word}+]")
+        elif marker == '?':
+            # Skip change marker lines
+            continue
+
+        # Wrap at 80 characters
+        if len(''.join(current_line)) > 80:
+            result.append(''.join(current_line))
+            current_line = []
+
+    if current_line:
+        result.append(''.join(current_line))
+
+    return '\n'.join(result)

transcript-fixer/scripts/utils/diff_formats/markdown_format.py

@@ -0,0 +1,104 @@
+#!/usr/bin/env python3
+"""
+Markdown report generator
+
+SINGLE RESPONSIBILITY: Generate detailed Markdown comparison report
+"""
+
+from __future__ import annotations
+
+from datetime import datetime
+from pathlib import Path
+
+from .change_extractor import extract_changes, generate_change_summary
+
+
+def generate_markdown_report(
+    original_file: str,
+    stage1_file: str,
+    stage2_file: str,
+    original: str,
+    stage1: str,
+    stage2: str
+) -> str:
+    """
+    Generate comprehensive Markdown comparison report
+
+    Args:
+        original_file: Original file path
+        stage1_file: Stage 1 file path
+        stage2_file: Stage 2 file path
+        original: Original text content
+        stage1: Stage 1 text content
+        stage2: Stage 2 text content
+
+    Returns:
+        Formatted Markdown report string
+    """
+    original_path = Path(original_file)
+    stage1_path = Path(stage1_file)
+    stage2_path = Path(stage2_file)
+
+    # Extract changes for each stage
+    changes_stage1 = extract_changes(original, stage1)
+    changes_stage2 = extract_changes(stage1, stage2)
+    changes_total = extract_changes(original, stage2)
+
+    # Generate summaries
+    summary_stage1 = generate_change_summary(changes_stage1)
+    summary_stage2 = generate_change_summary(changes_stage2)
+    summary_total = generate_change_summary(changes_total)
+
+    # Build report
+    report = f"""# 会议记录修复对比报告
+
+## 文件信息
+
+- **原始文件**: {original_path.name}
+- **阶段1修复**: {stage1_path.name}
+- **阶段2修复**: {stage2_path.name}
+- **生成时间**: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}
+
+## 修改统计
+
+| 阶段 | 修改数量 | 说明 |
+|------|---------|------|
+| 阶段1: 词典修复 | {len(changes_stage1)} | 基于预定义词典的批量替换 |
+| 阶段2: AI修复 | {len(changes_stage2)} | GLM-4.6智能纠错 |
+| **总计** | **{len(changes_total)}** | **原始→最终版本** |
+
+---
+
+# 阶段1: 词典修复详情
+
+{summary_stage1}
+
+---
+
+# 阶段2: AI智能修复详情
+
+{summary_stage2}
+
+---
+
+# 总体修改详情 (原始→最终)
+
+{summary_total}
+
+---
+
+## 使用说明
+
+1. **查看修改**: 每处修改都包含上下文,便于理解修改原因
+2. **人工审核**: 重点审核标记为"替换"的修改
+3. **专业术语**: 特别注意公司名、人名、技术术语的修改
+
+## 建议审核重点
+
+- [ ] 专业术语(具身智能、机器人等)
+- [ ] 人名和公司名
+- [ ] 数字(金额、时间等)
+- [ ] 上下文是否通顺
+"""
+
+    return report

transcript-fixer/scripts/utils/diff_formats/text_splitter.py

@@ -0,0 +1,33 @@
+#!/usr/bin/env python3
+"""
+Text splitter utility for word-level diff generation
+
+SINGLE RESPONSIBILITY: Split text into words while preserving structure
+"""
+
+from __future__ import annotations
+
+import re
+
+
+def split_into_words(text: str) -> list[str]:
+    """
+    Split text into words, preserving whitespace and punctuation
+
+    This enables word-level diff generation for Chinese and English text
+
+    Args:
+        text: Input text to split
+
+    Returns:
+        List of word tokens (Chinese words, English words, numbers, punctuation)
+    """
+    # Pattern: Chinese chars, English words, numbers, non-alphanumeric chars
+    pattern = r'[\u4e00-\u9fff]+|[a-zA-Z]+|[0-9]+|[^\u4e00-\u9fffa-zA-Z0-9]'
+    return re.findall(pattern, text)
+
+
+def read_file(file_path: str) -> str:
+    """Read file contents"""
+    with open(file_path, 'r', encoding='utf-8') as f:
+        return f.read()

transcript-fixer/scripts/utils/diff_formats/unified_format.py

@@ -0,0 +1,44 @@
+#!/usr/bin/env python3
+"""
+Unified diff format generator
+
+SINGLE RESPONSIBILITY: Generate unified diff format output
+"""
+
+from __future__ import annotations
+
+import difflib
+
+from .text_splitter import split_into_words
+
+
+def generate_unified_diff(
+    original: str,
+    fixed: str,
+    original_label: str = "原始版本",
+    fixed_label: str = "修复版本"
+) -> str:
+    """
+    Generate unified format diff report
+
+    Args:
+        original: Original text
+        fixed: Fixed text
+        original_label: Label for original version
+        fixed_label: Label for fixed version
+
+    Returns:
+        Unified diff format string
+    """
+    original_words = split_into_words(original)
+    fixed_words = split_into_words(fixed)
+
+    diff = difflib.unified_diff(
+        original_words,
+        fixed_words,
+        fromfile=original_label,
+        tofile=fixed_label,
+        lineterm=''
+    )
+
+    return '\n'.join(diff)