name: code-stats description: Analyze codebase statistics with tokei (line counts by language) and difft (semantic AST-aware diffs). Use for project composition, refactoring scope, codebase metrics. triggers: code statistics, line count, language breakdown, semantic diff, tokei, difft, codebase metrics, project size

Code Statistics

Rapid codebase analysis using two complementary tools.

Tools

tokei - Language Statistics

Line-count statistics across programming languages:

• File count per language
• Total lines, code lines, comments, blank lines
• Supports 150+ languages

Usage:

tokei                                    # Full project stats
tokei --type=TypeScript,JavaScript       # Filter by language
tokei --exclude="vendor/*"               # Exclude patterns
tokei -o json                            # JSON output for automation

difft - Semantic Diffs

AST-aware diffing that understands code structure:

• Ignores formatting changes
• Recognizes code moves and refactoring
• Shows meaningful changes, not line shuffles

Usage:

difft file1.go file2.go                  # Compare two files
GIT_EXTERNAL_DIFF=difft git diff         # Git integration
GIT_EXTERNAL_DIFF=difft git show HEAD    # Show last commit
GIT_EXTERNAL_DIFF=difft git log -p       # Log with semantic diffs

Workflow

1. Quick snapshot: tokei for project composition
2. Deep analysis: tokei -o json | jq for automation
3. Review changes: GIT_EXTERNAL_DIFF=difft git diff for meaningful diffs

MUST DO

• Use tokei before estimating project scope
• Use difft for refactoring reviews
• Filter output when only specific languages matter

MUST NOT

• Rely on raw line counts for complexity estimates
• Ignore comment ratios in health assessments

Advanced Features

• tokei advanced: See references/tokei-advanced.md for filtering, automation, CI patterns
• difft advanced: See references/difft-advanced.md for git integration, display modes, config