document-quality

name: document-quality description: Use when ensuring all generated documents meet quality standards including encoding, language, and formatting. Enforces Traditional Chinese and UTF-8 standards. metadata: version: "1.0.0"

Document Quality Skill

This skill ensures all research documents maintain consistent quality standards, particularly for encoding and language correctness.

When to use this skill

• Before committing any research documents
• After generating reports (daily briefs, ticker analyses)
• When encountering encoding errors
• During document review workflows

Quality Standards

1. Encoding Requirements

• UTF-8 without BOM (Byte Order Mark)
• No mixed encodings (Big5, CP950, GB18030)
• Consistent line endings (CRLF for Windows)

2. Language Requirements

• Traditional Chinese (繁體中文) for all content
• No Simplified Chinese (簡体中文)
• No晶晶體 (mixed Chinese-English), except:
  • Technical terms (API, Docker, Git)
  • Proper nouns (NVDA, BTC, TSMC)
  • Code snippets

3. Content Requirements

• No garbage/illegal characters
• Proper Markdown formatting
• Consistent spacing and punctuation
• Complete sentences (no truncated content)

Validation Tools

Automated Linter

.bin\python\python.exe ops/lint_reports.py

What it checks:

• UTF-8 encoding compliance
• Garbage/illegal characters
• Simplified Chinese character leaks
• Automatically fixes detected issues

Expected output:

--- SAR Report Linter V1.1 ---
------------------------------
Result: 77 Scan, 0 Issue Files, 0 Fixes.

Manual Review Workflow

Call /document-review

What it does:

• Comprehensive quality checklist
• Manual verification of critical content
• Final sign-off before commit

Common Issues & Solutions

Issue 1: Encoding Errors

Symptom: Garbled Chinese characters (亂碼)

Example: ?A?O?u??N?`??v

Solution:

# Auto-fix with linter
.bin\python\python.exe ops/lint_reports.py

# Manual conversion if needed
.bin\python\python.exe ops/convert_to_utf8.py <file_path>

Issue 2: Simplified Chinese Leaks

Symptom: Characters like ?体、优化、?据

Solution: Linter auto-detects and flags for manual correction

Issue 3: Garbage Characters

Symptom: Special characters like ?, ?, ?

Solution: Linter auto-removes or replaces with valid characters

Issue 4: Incomplete Content

Symptom: Truncated sentences, missing sections

Solution: Manual review required, regenerate report if necessary

Quality Checklist

Before committing any document, verify:

• Encoding: UTF-8 without BOM
• Language: Traditional Chinese only
• No晶晶體: Except technical terms
• No Garbage: Clean character set
• Formatting: Proper Markdown structure
• Completeness: All sections present
• Accuracy: Data and calculations correct

Auto-fix Capabilities

The linter can automatically fix:

Decision Tree

Q: When should I run quality checks?

• Always: Before Git commit
• Always: After generating reports
• Optional: During development (for quick feedback)

Q: Should I auto-fix or manual review?

• Auto-fix: Encoding, spacing, garbage characters
• Manual review: Simplified Chinese, incomplete content, critical errors

Q: What if linter reports issues?

1. Review the error output
2. Run auto-fix if available
3. Manually correct flagged issues
4. Re-run linter to verify
5. Commit only after clean result

Best Practices

1. Run linter early: Catch issues before they propagate
2. Commit frequently: Smaller changes are easier to review
3. Use UTF-8 editors: Ensure your editor saves as UTF-8
4. Validate inputs: Check data sources for encoding issues
5. Document exceptions: Note any intentional deviations from standards

Integration with Workflows

/daily-routine

• Phase 5: Document Review
• Automatically calls linter before commit

/ticker-deep-dive

• Section 7: Document Review
• Ensures ticker reports meet quality standards

/document-review

• Dedicated quality validation workflow
• Comprehensive checklist and sign-off