code-search-selector

Semantic search gates automatically route complex investigations through indexed search rather than bulk file reads. This prevents bias toward reading files directly and improves understanding across multiple files by prioritizing semantic relevance over proximity.

Use semantic search for audits, investigations, or tasks spanning three or more files. The gate detects these scenarios automatically and triggers before read operations, ensuring you search by meaning rather than file names or direct access patterns.

A stepwise decision process uses keyword-based triggers to detect investigation-type prompts, then routes queries to semantic indexing via claudemem. This mandatory gate ensures consistent, bias-free strategy selection before any file access occurs.

Semantic gates integrate with claudemem to index codebases and route queries to semantic search. The gate acts as a decision layer that determines when indexing is appropriate and directs queries to indexed search rather than native read tools.

Yes. By detecting when audits or broad investigations are needed, the gate prevents inefficient bulk reads and routes queries through semantic search, delivering faster and more accurate results for complex codebase analysis.