search-interceptor

Semantic search can replace bulk file reads by targeting specific concepts. When planning to read 3+ files or using broad glob patterns, use claudemem search to find relevant content efficiently, saving tokens and improving result quality.

Intercept bulk reads when accessing 3+ files in a directory, using patterns like **/*.ts or **/*.py, or coordinating parallel reads. The Interception Protocol pauses execution to evaluate indexing status, then executes semantic search instead of file reads.

The Interception Protocol pauses before bulk read execution, checks claudemem indexing status, and routes to semantic search if available. It provides fallback guidance for safe degraded flow if indexing is unavailable.

Yes. The Skill offers fallback prompts and error-handling templates for cases when indexing is unavailable, ensuring you have a safe degraded flow to proceed without semantic search optimization.

Semantic search targets specific concepts rather than reading entire files, reducing token consumption. By validating whether search can replace broad file reads, you improve signal quality and minimize unnecessary context overhead.