---

name: loctree description: Claude-specific loctree skill with PostToolUse hook augmentation for rg/grep results and explicit loct commands.

Loctree - Codebase Intelligence Skill (Claude)

Loctree provides AST-aware codebase understanding. Unlike text search (grep/ripgrep), loctree understands code structure: symbol definitions, exports, imports, dependencies, and dead code.

When This Skill Activates

This skill is automatically active through the PostToolUse hook. When you use Grep or Bash with rg/grep, loctree automatically augments results with semantic context.

You can also explicitly invoke loctree for deeper analysis.

Automatic Augmentation (Zero Friction)

When you search for code patterns, the hook automatically:

1. Detects symbol-like patterns (PascalCase, camelCase, snake_case)
2. Transforms regex to multi-query (A.*B → A|B)
3. Runs loct find to get semantic matches
4. Adds context: definitions, usages, cross-match files, dead code status

Example: You grep for ApiClient.*request → Hook transforms to request|ApiClient → Returns symbol definitions + semantic matches + cross-match files showing where both terms appear together.

Explicit Commands

loct find <symbol|pattern>

Search for symbol definitions across the codebase.

loct find "MyComponent"           # Single symbol
loct find "SymbolA|SymbolB"       # Multi-query (OR)
loct find "canonical_model|vision" # Find related symbols

Output includes:

• Symbol matches (definitions)
• Semantic matches (similar names)
• Parameter matches (function args)
• Dead code status

loct impact <file>

Analyze what would break if you change a file.

loct impact src/utils/helpers.ts

Use before refactoring to understand ripple effects.

loct slice <file>

Get the minimal dependency slice for a file.

loct slice src/components/Button.tsx

Shows only the files needed to understand this file.

loct health

Codebase health report: dead code, cycles, duplicates.

loct health

Output includes:

• Dead exports (unused code)
• Circular dependencies
• Duplicate exports (twins)
• Health score (0-100)

loct focus <directory>

Analyze a specific directory's structure.

loct focus src/api/

loct query who-imports <file>

Find all files that import a given file.

loct query who-imports src/hooks/useAuth.ts

loct --for-ai

Get a comprehensive codebase overview optimized for AI understanding.

loct --for-ai

Run this first when starting work on a new codebase.

Best Practices

1. Start with loct --for-ai to understand codebase structure
2. Use loct find instead of grep for symbol searches (10x better results)
3. Check loct impact before refactoring to avoid breaking changes
4. Run loct health periodically to catch dead code and cycles

Pattern Recognition

The hook automatically recognizes these patterns:

Pattern Type	Example	Action
PascalCase	MyComponent	loct find
camelCase	handleClick	loct find
snake_case	user_service	loct find or Tauri command
File path	src/utils.ts	loct slice
Directory	src/api/	loct focus
Multi-term	A|B or A.*B	loct find "A|B"
Health keywords	dead, unused, cycle	loct health

Cross-Match Files

When searching multiple terms, loctree identifies files containing 2+ different terms - these are often the most relevant files for understanding relationships.

Prerequisites

Install loctree CLI:

cargo install loctree
# or
brew install loctree

Verify installation:

loct --version

Initialize in your project (creates .loctree/ cache):

loct scan

---

Made with (งಠ_ಠ)ง by the ⌜ Loctree ⌟ 𝖙𝖊𝖆𝖒 (c) 2026 Maciej & Monika + Klaudiusz (AI)