scrutinize

To perform deep code review for logic correctness and security, run a script that detects off-by-one errors, null paths, race conditions, and SQL injection vulnerabilities. This requires type checking, linting, and build steps to be completed prior to execution.

To audit edge cases and security in a Supabase codebase, use a review process that identifies type coercion, JSONB casting issues, RLS bypass, exposed secrets, and unvalidated input to ensure data flow shape mismatches are caught.

Checking for dead code and orphaned exports before merging involves running an analysis script to identify unreachable branches and orphaned exports. This ensures high-quality, secure, and maintainable codebases by validating import direction compliance and removing dead code.

Yes, deep code review requires type checking, linting, and build steps to be completed prior to execution. This prerequisite ensures the analysis can focus on in-depth logic correctness, security vulnerabilities, and edge case handling beyond basic syntax checks.

You should use a deep code analysis script instead of basic syntax checks to detect complex issues like asynchronous race conditions, retry storms, and network failure paths. It provides in-depth analysis for logic correctness, security, and edge cases beyond the scope of policy-based gates.