Language focus
AI code review for JavaScript
JavaScript spreads across browsers, servers, bundles, and build chains. Automated review works best when the model sees enough context - a full module or a coherent chunk - not three lines torn out of framework glue.
Where CodeCritic concentrates for JavaScript
- Async flow, stale closures, and missing error boundaries that unit tests rarely catch in isolation.
- Package boundaries (CommonJS versus ESM) and brittle default exports spread through large codebases.
- Security hotspots: dangerouslySetInnerHTML, eval-shaped patterns, JWT handling mistakes, permissive CSP gaps.
- Lint-friendly refactors versus behavior changes - CodeCritic calls out edits that silently shift semantics.
AI review does not substitute for threat modeling or compliance sign-off - treat it as a fast second pass alongside human owners who know deployment risk.