Run parallel multi-model code reviews with Claude Code subagents and external LLMs.
npx skills add https://github.com/zuozuo/super-review --skill super-review
Please help me install this Agent Skill. Skill: super-review Source: https://github.com/zuozuo/super-review/tree/main/super-review Command: npx skills add https://github.com/zuozuo/super-review --skill super-review
This Skill enables parallel, multi-model code review using Claude Code subagents and external LLMs to provide faster, more thorough insights across reviewers.
Run super-review on a target file, commit, or directory to start the parallel review workflow.
High-intent search queries and answers about installing and using this skill.
To run a multi-model code review, trigger the workflow with a prompt like 'super-review' and provide a diff, commit, or file path. Claude Code subagents and external LLMs concurrently inspect changes to generate structured reports.
The best way to consolidate AI code review findings is to deduplicate results and categorize them by severity. This approach unifies verdicts across multiple reviewers and generates a consolidated summary focusing on critical P0 and P1 issues.
Yes, you can use Claude Code subagents with external LLMs like Codex and Gemini CLI. This combination concurrently runs multiple reviewers to inspect changes, delivering faster and more thorough insights across the board.
Multi-model code review handles scoring by automatically selecting evidence-based prompt templates to evaluate compliance, UI, and quality. It generates comprehensive scores and stores per-reviewer metadata in a dedicated .reviews directory.
Yes, parallel code review works with specific git commits and file paths. You can target a specific file, commit, or directory to start the concurrent review workflow, which processes these inputs to produce structured reports.
Focusing multi-model code reviews on P0 and P1 severity issues accelerates fixes and release readiness. By deduplicating findings and categorizing critical problems, developers can quickly address major compliance, UI, and quality defects.
zangjeicyReview recent git diffs and apply safe code cleanup fixes.
seanchatmangptRun adversarial reviews on PRs, documentation, or architecture with three review levels.
louloulinAutomate CI/CD pipelines and infrastructure provisioning across Jenkins, GitLab, and GitHub Actions.
danielgapRuns two parallel LLM reviewers to produce a fused verdict report.
