ce:review

Coordinate multi-agent code review workflows and generate structured findings.

Updated Apr 22, 2026
One-click install
npx skills add https://github.com/yxc023/agent-config-studio --skill ce-review-yxc023
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ce:review
Source: https://github.com/yxc023/agent-config-studio/tree/main/.opencode/skills/ce-review
Command: npx skills add https://github.com/yxc023/agent-config-studio --skill ce-review-yxc023

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

ce:review orchestrates multi-agent code reviews by coordinating tiered personas to analyze diffs, produce a structured findings report, and drive a merge/dedup pipeline with run artifacts for downstream workflows.

Core Features & Use Cases

  • Always-on reviewers: correctness, testing, maintainability, and project-standards, plus CE agents agent-native-reviewer and learnings-researcher.
  • Conditional reviewers: security, performance, api-contract, data-migrations, reliability, adversarial, cli-readiness, and previous-comments, selected based on diff content.
  • Structured outputs: findings conform to a formal schema with residual risks, testing gaps, and clear ownership/auto-fix routing.
  • Run artifacts and headless outputs: supports automated pipelines with per-run directories and JSON artifacts.
  • Standalone and PR reviews: intent discovery, plan discovery, and staged synthesis integrated into the compound-engineering workflow.

Quick Start

Run a ce:review session on a changed codebase to generate a structured report and associated run artifacts.

Frequently Asked Questions about ce:review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate structured code reviews for diffs in a CI pipeline?

Automated code review routes diff findings through tiered agents to generate structured reports. It classifies results into safe_auto, gated_auto, manual, and advisory categories, producing JSON artifacts for downstream pipeline consumption.

What is the best way to run multi-agent code reviews with specialized tiered personas?

Multi-agent code review coordinates always-on reviewers for correctness and testing alongside conditional agents for security and performance. This routes findings through an autofix pipeline, creating actionable next steps based on diff content.

Can I use automated code review for both standalone analysis and pull requests?

Automated code review supports both standalone and pull request workflows. It integrates intent discovery, plan discovery, and staged synthesis to produce structured findings with clear ownership and auto-fix routing.

How does multi-agent code review classify and route findings for autofix pipelines?

Multi-agent code review classifies findings into safe_auto, gated_auto, manual, and advisory tiers. This structured routing ensures residual risks and testing gaps are addressed with clear ownership for automated remediation.

Do I need to provide a run_id to generate artifacts during automated code review?

Providing a run_id generates per-run artifacts and headless JSON outputs. While the structured findings report is always produced, the run_id enables integration with automated compound-engineering workflows.

What limitations exist when selecting conditional reviewers for specific diff content?

Conditional reviewers for security, performance, api-contract, data-migrations, reliability, adversarial, cli-readiness, and previous-comments are selected based on diff content. The selection is automatic, limiting manual override of the reviewer composition.