How does this skill save tokens?

It guides Claude to use low-token tools like AST-grep (~50 tokens) for initial discovery instead of high-token 'Read' commands (1500+ tokens) which should only be used as a last resort.

What is the benefit of using LEANN with this skill?

LEANN is prioritized for semantic queries like 'how does auth work,' providing conceptual answers and relevant paths with minimal token overhead compared to raw Grep.

What happens if Claude tries to use an inefficient tool?

The integrated hook automatically detects the query type, blocks the inefficient command, and suggests the optimal tool (like AST-grep for structural queries) to maintain efficiency.

Does this replace standard Grep functions?

No, it organizes the workflow so Grep is used only for its strengths—literal identifier matches and regex—while structural and semantic queries are routed to more efficient tools.

Maestro Search Hierarchy

Name: Maestro Search Hierarchy
Author: scooter-lacroix

byscooter-lacroix

生産性とワークフロー

Optimizes Claude's code search strategy to minimize token consumption and improve query accuracy.

概要

The Maestro Search Hierarchy skill implements a token-efficient decision tree for code exploration, guiding Claude to use the most appropriate tool for every query. By prioritizing AST-grep for structural patterns, LEANN for semantic concepts, and standard Grep only for literal identifiers, it drastically reduces unnecessary context loading. This skill includes built-in enforcement hooks that prevent expensive 'Read' operations before relevant files are identified, ensuring faster responses and significantly lower API costs during complex development tasks.

主な機能

Automatic hook enforcement via grep-to-leann scripts
Semantic search via LEANN for conceptual discovery
Token-efficient decision tree for code navigation
AST-grep integration for structural code patterns
0 GitHub stars
Significant reduction in context window usage

ユースケース

Finding specific function or class definitions across large repositories
Locating exact string identifiers or regex patterns efficiently
Understanding high-level architecture and patterns without reading every file

概要

主な機能

Automatic hook enforcement via grep-to-leann scripts
Semantic search via LEANN for conceptual discovery
Token-efficient decision tree for code navigation
AST-grep integration for structural code patterns
0 GitHub stars
Significant reduction in context window usage

ユースケース

Finding specific function or class definitions across large repositories
Locating exact string identifiers or regex patterns efficiently
Understanding high-level architecture and patterns without reading every file