Critic by Feyn

AI coding agents can produce changes faster than people can review them, and a diff alone does not say why a change was made. Critic, listed here as Critic by Feyn, tries to put the reasoning back in front of whoever checks the change.

Change write-ups from the agent itself

Critic is added to a coding agent as a plug-in, and from then on the agent documents each change while making it. The account covers what changed, the assumptions made, the decisions taken, the complex parts, and evidence such as screenshots and instructions for running the code. Whoever checks the change reads this account beside the code. Feyn showed the product on 24 September 2026, and the site footer names Chonkie, Inc.

Questioning the original session

The person checking can put questions to a fork of the original agent session, which answers them. That means questioning the very context that produced the change, not a fresh model guessing at it. All of it can also be reached over MCP, and changes pushed to GitHub are mirrored under GitHub's own permissions. The demonstration on the site shows a change page with the branch, the people involved, the agents and a threaded discussion on the diff.

What to ask before adopting it

Review becomes the bottleneck for teams letting agents write a growing share of their code, which is the problem Critic addresses. A team should ask before adopting it where the session of the agent and the code are kept, and who has access to them. No price is published and the terms name no company. The Critic by Feyn name is used because several code-review tools have similar names.