file-search

The file-search skill is a high-performance utility designed for developers and AI agents to efficiently navigate, audit, and analyze complex codebases. By providing a unified interface for both pattern-based text search and syntax-aware structural matching, it significantly reduces the time required to locate entry points, identify usage patterns, or detect security vulnerabilities. This skill is built on the performance-oriented engines ripgrep and ast-grep, which are optimized for rapid indexing and retrieval across large-scale projects, making them superior to standard grep or find commands.

• Utilize ripgrep (rg) for lightning-fast text-based searching of variable names, strings, comments, and regex patterns across multiple files and directories.

• Leverage ast-grep (sg) for advanced, syntax-aware structural searching, allowing for matching against abstract syntax trees to find function signatures, class definitions, or complex code blocks irrespective of formatting.

• Enable rapid codebase discovery by finding TODOs, FIXMEs, specific API calls, or hardcoded credentials during security audits.

• Enhance refactoring workflows by identifying all cross-references to a specific class, method, or dependency before making structural changes.

• Support large-scale code analysis by filtering results using language-specific flags, glob patterns for excluding directories, and directory-specific scoping to maintain a high signal-to-noise ratio.

• Implement professional search workflows, such as counting match occurrences, piping outputs to head for sampled inspection, and utilizing word boundaries for precise result matching.

• Always start searches with targeted, narrow scopes to prevent overwhelming the context window and wasting tokens.

• Recommended usage involves a progressive refinement approach: begin by identifying the general location, then narrow the scope by file type (e.g., --type py, --lang js) and directory before applying complex regex or structural patterns.

• When performing complex refactoring, prioritize ast-grep's capability to understand code semantics over standard text-based tools to ensure that matches are functionally relevant rather than just string matches.

• Always be mindful of token consumption; pipe large result sets to head or count them first to verify the search density before requesting full content access.

• Ensure the environment remains isolated; this skill is specifically optimized for use within the MassGen execution framework, leveraging pre-configured binary tools available in the container environment.