agent-pr-replay

🥈Silver

Agent PR Replay takes merged PRs from any repository, reverse-engineers the task prompt, runs Claude Code against it, and compares what the agent did versus what humans actually shipped. The result is targeted, empirical guidance.

2660Updated 1w ago

Intermediate30min to implementautomation

Saves ~240 min per use

Quick InstallView Source

git clone https://github.com/sshh12/agent-pr-replay.git

Works with:

Claude

Overview

About This Skill

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/sshh12/agent-pr-replay

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Analyze the following merged PR from [REPOSITORY] in [COMPANY]'s [INDUSTRY] codebase. Reverse-engineer the task prompt, run Claude Code against it, and compare the AI's output with the human implementation. Provide targeted, empirical guidance on how the AI could improve or where the human approach was superior. Here is the PR: [PR_DATA].

Example Output

## Analysis of PR #1234: User Authentication Refactor

### Reverse-Engineered Task Prompt
- Implement OAuth 2.0 authentication flow
- Refactor existing session management
- Add rate limiting to prevent brute force attacks
- Ensure backward compatibility with legacy systems

### Claude Code Output vs Human Implementation

**AI Implementation**
- Used JWT for session management
- Implemented rate limiting with Redis
- Added comprehensive unit tests

**Human Implementation**
- Kept existing cookie-based sessions for legacy support
- Implemented rate limiting with database-backed counters
- Added integration tests for OAuth flow

### Empirical Guidance

**Where AI Excelled**
- The JWT implementation was more secure and scalable than the human approach
- AI's unit test coverage was more comprehensive

**Where Human Approach Was Superior**
- Legacy system compatibility was critical for [COMPANY]'s [INDUSTRY] needs
- Database-backed rate limiting was more maintainable for the team's current infrastructure

**Recommendations**
- Consider a phased migration to JWT for session management
- Adopt AI's unit test approach for future features
- Document rate limiting implementation choices for future reference