code-audit

A parallel code audit coordinates multiple specialized reviewers scanning your codebase simultaneously for security, reliability, and quality validation. This skill orchestrates that team, discovers the project structure, and merges findings into a unified backlog.

Yes, you can automatically create Linear issues from code review findings. The audit validates discovered defects and publishes them directly as Linear issues in the backlog, complete with tasks, statuses, and labels for triage and remediation.

Yes, the code audit works with TypeScript and Node.js projects by performing pre-flight discovery of the project structure via tsconfig.json and package.json. It applies CLAUDE.md rules to validate the codebase across varying project sizes.

If agent team creation fails during a code audit, the skill provides automatic fallback to a single-agent audit. This ensures the codebase is still reviewed for security, reliability, and quality defects without interruption.

The best way to audit a large codebase for security vulnerabilities is orchestrating domain-focused parallel reviews. This skill assigns dedicated security, reliability, and quality reviewers to scan the code and merge results into a Linear backlog for remediation.