dev-review

Organize browser-based developer review gates before merge decisions.

1|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/SeoJaeWan/try-claude-code --skill dev-review-seojaewan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dev-review
Source: https://github.com/SeoJaeWan/try-claude-code/tree/main/plugin/develop/skills/dev-review
Command: npx skills add https://github.com/SeoJaeWan/try-claude-code --skill dev-review-seojaewan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Organizes browser-based developer review gates after runner completes all phase commits and before merge decisions. It collects per-card feedback and routes non-approved changes back into the runner's worktree for rework, Q&A, or out-of-scope records.

Core Features & Use Cases

  • Provides a browser-based UI that walks Overview → per-commit cards → Final preview, capturing per-card status (approved / needs-change / question / out-of-scope) and an optional dispatch_agent for rework.
  • Maintains artifacts under plans/{task_slug}/dev-review/ (review-data.json, feedback.json, review-history.json) to support multi-round review cycles and auditability.
  • Re-generates deterministic review data using a helper script and serves a local static UI, preserving previously approved cards when rounds continue and enabling iterative feedback loops.

Quick Start

Open the dev-review UI served locally and complete the review, then reply with '리뷰 완료' to proceed.

Frequently Asked Questions about dev-review

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

FAQPage Schema
How do I automate code review gates before merge decisions?

Automate code review gates by serving a local static UI that organizes per-commit cards after phase commits. This browser-based workflow captures per-card feedback and routes non-approved changes back into the worktree for rework.

Can I persist iterative code review history for multi-round feedback cycles?

Yes, persist iterative code review history by storing artifacts like review-data.json, feedback.json, and review-history.json under plans/{task_slug}/dev-review. This supports multi-round review cycles and auditability while preserving previously approved cards.

What is a browser-based developer review workflow for commit changes?

A browser-based developer review workflow walks reviewers through Overview to per-commit cards to a Final preview. It captures statuses like approved, needs-change, question, or out-of-scope, and optionally dispatches an agent for rework.

Do I need a deterministic review-data generator to run a local review UI?

Yes, you need a deterministic review-data generator and a static server to serve the UI. The workflow also requires well-formed plan_signature, task_head_sha, and available_agents to properly drive the automated review process.

How to handle out-of-scope or needs-change feedback during automated code review?

Handle out-of-scope and needs-change feedback by capturing the status per commit card in the browser UI. The workflow then routes these non-approved changes back into the runner's worktree for rework, Q&A, or out-of-scope records.

Does dev-review preserve approved cards when continuing iterative review rounds?

Yes, dev-review preserves previously approved cards when rounds continue. It re-generates deterministic review data using a helper script and serves the local static UI, enabling iterative feedback loops without losing prior approvals.