Coordinate subagents to review diffs, commits, and PRs with severity-filtered findings.
npx skills add https://github.com/joesys/joesys-skills --skill code-review-joesys
Please help me install this Agent Skill. Skill: code-review Source: https://github.com/joesys/joesys-skills/tree/main/skills/code-review Command: npx skills add https://github.com/joesys/joesys-skills --skill code-review-joesys
Analyze codebases automatically to surface correctness, code quality, architecture, reliability, security, and performance violations, delivering structured findings with concrete before/after guidance.
Invoke /code-review on the targeted code changes to generate a comprehensive, prioritized review report.
High-intent search queries and answers about installing and using this skill.
Automate code reviews for large pull requests by using a phase-driven workflow that resolves scope, performs parallel subagent analysis across domains, and synthesizes prioritized findings. This handles varying diff sizes without manual line-by-line inspection.
Parallel domain analysis in automated code review evaluates seven focus areas concurrently: correctness, clean code, architecture, reliability, security, performance, and readability. This mechanism surfaces comprehensive violations across multiple disciplines simultaneously.
Deduplicate and filter code violations by severity through cross-model aggregation and corroboration. The process removes overlapping findings from parallel analyses and outputs a confidence-weighted synthesis with only prioritized, actionable results.
Automated code review applies to individual commits and small diffs through a scoped phase-driven workflow. It resolves the change scope first, then executes targeted per-cluster or per-file analysis to provide structured findings.
Cross-model code review provides concrete before and after fix guidance. By aggregating and corroborating findings across multiple models, it delivers actionable synthesis with ready-to-apply code corrections for identified architectural and security issues.
The best way to review codebase architecture and security violations is dispatching parallel domain-specific subagents. This approach applies targeted analysis to specific focus areas, ensuring comprehensive coverage of reliability and performance issues.
vercel-labsAnalyze code changes and provide actionable feedback on defects, style, and design.
oimiragieoIdentify violated code rules and provide step-by-step fix instructions.
AnveioAutomate GitHub pull request creation, reviews, feedback, and merges via gh CLI.
thomfilgAutomate Jira task creation through multi-agent consensus with parallel and sequential iterations.
