How are the findings prioritized?

The skill maps architectural severity to task priority: Blocker and High findings become p1 tasks, Medium findings become p2, and Low findings become p3.

Where are the generated todos stored?

All findings (except nits) are converted into structured markdown files and stored in the .claude/todos/ directory of your project for easy tracking.

How does the parallel review process work?

The skill spawns four independent Task agents simultaneously, each focusing on a specific domain: core architecture, performance, scalability, and API contracts, ensuring comprehensive coverage in less time.

Can this skill analyze uncommitted changes?

Yes, it specifically targets the current working tree changes using git diff, making it ideal for pre-commit or pre-push reviews.

Architectural Code Review & Task Automation

Name: Architectural Code Review & Task Automation
Author: jayteealao

byjayteealao

0•

コラボレーションツール

Performs multi-dimensional parallel architecture reviews and automatically converts findings into prioritized project todos.

The Architecture Code Review skill orchestrates four specialized Task agents to analyze code changes across architecture, performance, scalability, and API contracts simultaneously. By examining git diffs, it generates a unified report that deduplicates findings and sorts them by severity. Beyond mere analysis, this skill bridges the gap between review and remediation by automatically populating your project's .claude/todos/ directory with structured markdown tasks, ensuring that architectural debt and critical bottlenecks are tracked and resolved within your existing developer workflow.

主な機能

010 GitHub stars

02Severity-based prioritization (Blocker to Nit) for actionable feedback

03Parallel execution of specialized architecture, performance, and scalability agents

04Automatic generation of markdown-based todo files with metadata and acceptance criteria

05Automated merging and deduplication of findings into a unified health assessment

06Context-aware analysis using git diffs to focus on recent working tree changes

ユースケース

01Conducting deep-dive architectural audits before merging major refactors

02Automating the documentation of technical debt and required fixes during code reviews

03Ensuring API contract stability and system scalability in high-growth codebases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jayteealao/agent-skills review-architecture

For use in Claude.ai and ChatGPT

Download Skill