reviewing-code

Code reviews coordinate 12 parallel agents—six from pr-review-toolkit and six custom—to evaluate PRs, unpushed commits, or uncommitted work. Trigger via /code:review-pr, /code:review-unpushed, or /code:review-uncommited to deploy agents covering security, performance, quality, consistency, design, and governance in a single comprehensive report.

Parallel code reviews evaluate security vulnerabilities, performance bottlenecks, code quality issues, consistency violations, design problems, and governance compliance. Each of the 12 agents reports findings with exact file paths and line numbers, ensuring no issues are missed and all results follow a standardized format.

Yes. The code review workflow supports three triggers: /code:review-pr for pull requests, /code:review-unpushed for commits not yet pushed, and /code:review-uncommited for local changes. All three activate the same 12-agent parallel execution to produce identical reporting standards.

All 12 agents report findings regardless of outcome. The system implements startup rules that determine whether each agent runs or skips, then outputs pre-launch status and final results in a standardized format, ensuring transparency and no silent failures.

Results preserve per-agent issues with exact file paths and line numbers in a standardized format. Each of the 12 agents reports independently, allowing you to see which checks passed, which flagged problems, and where in your code changes those problems occur.

No dependencies or components are required. The code review workflow is self-contained and runs immediately when triggered. Simply invoke /code:review-pr, /code:review-unpushed, or /code:review-uncommited to activate the 12-agent parallel review system.