capsa

SQL validation for teams that want bad queries caught before they ship.

capsa parses SQL, validates dialect syntax, and scans the query against rule sets for security, performance, reliability, and data-handling risk. It does not execute SQL. It gives you fast feedback on the query you are about to run, review, or deploy.

What It Does

• Parses SQL into statements for validation
• Supports dialect-aware validation through parser plugins
• Checks query against compiled rules
• Runs schema validation and rule validation in parallel
• Returns a compact list of findings

Usage

1. Build: cargo build --profile release -p cli
2. Run: capsa -q "SELECT * FROM users" - ouputs: ["SELECT *; list required columns", "Unbounded SELECT; add WHERE, LIMIT, or pagination"]
3. Dialect (optional): capsa -q "SELECT * FROM users" -d postgres. Supported dialects: ansi, bigquery, clickhouse, databricks, duckdb, generic, hive, mssql, mysql, postgres, redshift, snowflake, sqlite.

Build

Dialects are implemented as plugins. To build with a specific dialect: cargo build -p core-sql --features postgres.