I built this after noticing that AI-agent PRs are hard to review. There are many tools available but it does not show the reason why the code can be problematic in case of agent. A normal diff can show code changes, but it may not make it obvious that a prompt changed, a model changed, a tool c...

Source: [Hacker News](https://news.ycombinator.com/item?id=49067759)

Sponsored