codebase_search

The codebase_search utility facilitates semantic exploration. Rather than relying on rigid text matching, it interprets the conceptual intent of your query to locate relevant logic and structures across your entire project history.

[!NOTE] This capability requires an active and synchronized Codebase Index.

Parameters

• query (required): A natural language description of your target (e.g., "how is user session persistence implemented?").
• path (optional): Focus the search on a specific architectural layer (e.g., lib/security).

The Advantages of Semantic Retrieval

Standard search tools (grep) often fail if you aren't using the exact variable name. codebase_search overcomes this by identifying:

• Conceptual Relationships: Inputting "payment processing" could surface executeStripeTransaction() and BillingService.js, even if the exact phrase "payment processing" is absent.
• Structural Patterns: Locate architectural styles or error-handling strategies that share a similar conceptual signature.

Relevance Scoring

Search results are prioritized using a similarity score (ranging from 0.0 to 1.0). To ensure the highest accuracy, Sypha typically filters for matches with a confidence score above 0.4.

Operational Tips

• Provide Detail: Vague terms like "database" are less effective than "establishing a connection pool to PostgreSQL via TypeORM."
• Limit Context: If you are certain a feature resides in the frontend, providing a path parameter like "apps/web" increases both speed and precision.