structural-search

Structural search uses AST (Abstract Syntax Tree) patterns to identify code by its structure rather than regex. ast-grep enables this by parsing code into syntax trees, then matching patterns like function calls, imports, and class definitions across multi-language codebases without false positives from text-based searches.

Yes. ast-grep identifies function calls by their structural definition in the AST, allowing you to find exact usages without regex noise. You define patterns using syntax tokens like $NAME and $_ to match function calls, then search across your entire codebase to locate all occurrences.

Structural search identifies refactoring targets by matching anti-patterns and code structures that need updating. ast-grep lets you define patterns for deprecated APIs, unsafe patterns, or style violations, then find all instances across multiple files and languages for systematic refactoring.

Yes. ast-grep parses code into language-agnostic AST structures, so a single pattern definition can match equivalent code structures across different languages. This enables you to search for similar patterns in polyglot codebases without writing language-specific regex.

After identifying patterns with ast-grep's search functionality, you can apply replacements using the same pattern syntax. The Skill supports replace operations through Bash workflows, letting you update matched code structures and export results in JSON or YAML format for validation.

Structural search requires ast-grep tooling and pattern syntax learning, but eliminates false positives that regex causes. The trade-off is setup complexity for precision: you gain exact semantic matching at the cost of regex's simplicity, making it best for large codebases where false positives are costly.