ce-review

Automate multi-agent code review by spawning targeted reviewer personas to analyze PR diffs and merge their findings into a single structured report. The process handles intent discovery, plan verification, and reviewer selection automatically across diverse tech stacks.

Intent discovery and diff analysis on branch changes are best handled by spawning specialized sub-agents that verify implementation plans against the actual diff. These agents preserve the diff context without mutating prompts, ensuring accurate artifact synthesis across the reviewed branch.

Automated PR review supports standalone checks on individual commits without requiring full branch context. The multi-agent system applies always-on personas for correctness, testing, maintainability, and project standards to synthesize findings from the isolated commit diff.

Automated code review works across diverse tech stacks by deploying cross-cutting and stack-specific reviewer personas. It evaluates project standards alongside correctness and maintainability to ensure the synthesized report accurately reflects the unique requirements of the technology stack.

Generate a structured code review report by merging the findings from multiple specialized reviewer sub-agents into a single synthesized artifact. This multi-agent coordination evaluates the diff scope and produces a consolidated report covering correctness, testing, maintainability, and standards.