quality-review

You can automate AGENTS.md code reviews by using a skill that parses the Implementation Quality section and dispatches parallel agents per criterion to evaluate changes. It triages raw findings and compiles a final summary table, providing structured feedback aligned to your project standards.

The best way to triage code review findings is to use an automated skill that filters raw results to drop false positives, duplicates, and contradictions. This skill performs that triage automatically after parallel agents evaluate your changes, compiling a clean, summarized report.

Yes, an AGENTS.md file is required to run automated implementation quality reviews. The skill specifically parses the Implementation Quality bullets from your repository's AGENTS.md file to map each bullet to an individual review criterion for the parallel agent evaluation.

Multi-agent dispatch for quality review works by sending a single message that assigns a dedicated Task agent to each criterion in parallel. This maximizes efficiency by evaluating all AGENTS.md Implementation Quality criteria simultaneously before the results are triaged and compiled.

Yes, you can use distributed agents to evaluate code changes for large projects. This skill dispatches a dedicated agent per criterion in parallel within a single message, allowing multiple AGENTS.md quality checks to run simultaneously and efficiently scale the review process.