review-fix-loop

A fix-and-review loop uses parallel subagent investigation and iterative reviews with quality gate checks to manage complex tasks. This mechanism requires a structured prompt with clear goals and scope to ensure continuous improvement and high-quality outcomes.

Automating code review is best handled by parallel subagent investigation combined with iterative review scoring. This method uses evidence-based triage and quality gate checks to evaluate task quality continuously throughout the software development lifecycle.

Parallel subagent investigation works by leveraging multiple subagents to perform parallel tasks simultaneously. This reduces overall time to completion for complex software development scenarios by distributing the investigation workload before iterative reviews begin.

Iterative review scoring requires a structured prompt with clear goals and scope to function properly. Without this structured input, the fix-and-review loop cannot effectively perform evidence-based triage or evaluate quality gate checks for complex tasks.

Quality gate checks are needed when complex software development tasks require multiple iterations to achieve desired quality. They ensure high-quality outcomes by applying evidence-based triage and review scoring before completing the fix-and-review loop.