Back to search
Top Contributors
8
commits
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.
Stars
32
Forks
7
Watchers
32
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
8
commits
Add 'Show me the prompts!' section with permalinks to all LLM prompts
3904559View on GitHubAdd agent-heatmap CLI tool for analyzing AI coding agent behavior
4335b1dView on GitHub