skills.rest
ybbms777
ybbms777COMMUNITY
Explore all skills by this author →
GitHub

ce:review

Execute parallel code review with tiered reviewer personas for logic, security, and performance issues.

1|Updated Apr 4, 2026
Search Tags:#code quality#code review#pr review#security review#automated review#defect detection#persona agents
One-click install
npx skills add https://github.com/ybbms777/compound-engineering --skill ce-review-ybbms777
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ce:review
Source: https://github.com/ybbms777/compound-engineering/tree/main/skills/ce-review
Command: npx skills add https://github.com/ybbms777/compound-engineering --skill ce-review-ybbms777

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Manual code reviews are often inconsistent, miss critical issues across specialized domains like security and performance, and produce duplicated or unprioritized feedback that slows down PR merges. This skill automates comprehensive code review by spawning parallel, specialized reviewer personas, merging and deduplicating their findings, and prioritizing issues by severity and actionability to streamline the PR review process.

Core Features & Use Cases

  • Tiered Parallel Reviewer Personas: Automatically spawns always-on reviewers for correctness, testing, maintainability, and project standards, plus conditional reviewers for security, performance, API contracts, and stack-specific patterns based on the diff content.
  • Confidence-Gated Finding Prioritization: All findings are scored by confidence and severity (P0-P3), with clear action routing for safe auto-fixes, gated changes requiring approval, manual downstream work, and advisory notes.
  • Flexible Operation Modes: Supports interactive review with user prompts, fully autonomous autofix mode, read-only report-only mode for parallel verification, and headless mode for programmatic skill-to-skill invocation.
  • Use Cases: Use before creating a PR, after completing iterative implementation tasks, when requesting feedback on code changes, or as a read-only verification step in larger engineering workflows.

Quick Start

Invoke the ce:review skill with no arguments to review all changes on your current working branch before opening a pull request, or pass a PR link or branch name to review targeted changes from an existing pull request or separate branch.

Frequently Asked Questions about ce:review

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

FAQPage Schema
How do I automate code review before creating a pull request?▼

Automated code review before a pull request is handled by invoking the skill with no arguments to scan all working branch changes. It uses parallel persona agents to identify logic errors, security vulnerabilities, and performance regressions.

Can I run a security review and performance check on my code diff simultaneously?▼

Security review and performance checks can run simultaneously via tiered parallel reviewer personas. Conditional reviewers for security and performance are automatically spawned based on the specific content of the code diff.

What is the best way to prioritize code defects found during PR review?▼

Prioritizing code defects during PR review is achieved through confidence-gated finding prioritization. All identified issues are scored by confidence and severity from P0 to P3, routing actions for auto-fixes, gated approvals, or manual work.

Does automated code review work in a headless mode for programmatic workflows?▼

Automated code review supports headless mode for programmatic skill-to-skill invocation. It also offers interactive, fully autonomous autofix, and read-only report-only modes to fit different engineering workflows.

How do I review an existing pull request from a separate branch?▼

To review an existing pull request, pass the PR link or branch name as an argument to the skill. It will execute structured code review on the targeted changes and merge deduplicated findings.

What code quality issues can specialized reviewer agents detect in my changes?▼

Specialized reviewer agents detect logic errors, security vulnerabilities, performance regressions, and maintainability flaws. Always-on reviewers check correctness and testing, while conditional reviewers handle stack-specific patterns.

Related Skills

Explore More Software Engineering Skills→
antoniocascaisantoniocascais
11

pr-review

Analyze git diffs to assess code quality, security, and merge readiness.

Community
Advanced
shogoggshogogg
4

coderabbit-review

Automate CodeRabbit code reviews for uncommitted and committed Git changes.

Community
Advanced
mattsp1290mattsp1290

pr-ready

Automate code review, security verification, and merge readiness checks for software projects.

Community
Advanced
Next-ofkinNext-ofkin
1

code-review-ai-ai-review

Automate code review with static analysis and AI models.

Community
Advanced