"Co-Authored-By: Claude" is fine

Posted on: Mar 07, 2026

As part of Git, you can add the Co-Authored-By trailer to a git commit mesage to include multiple people in a single commit. As part of this, if you use Co-Authored-By: Claude <noreply@anthropic.com> it will have Claude appear as a contributor on places like GitHub. I've seen a lot of sentiment on internet forums of mixed opinion on whether this is a good addition or just "marketing" / slop.

Note, Claude Code injects these attributions (Claude Code docs ref) by default if you have it run git commit itself, but that is not something I particularly do myself as I try to still be in the drivers seat and manage git myself rather than let Claude or other AI agents do it.

I personally try to add the attribution when I am in a coding session where Claude or other agents are doing a portion of the work of either generating or refactoring code, as there is something more to this more than say using something like OpenAPI or Java annotations to do code generation. The mentality I am at least having currently is I would prefer to be open that I was using these tools in part of the process of my software development rather than try to be sneaky and hide that I am using them. As well, during this process I am still reading the code before doing a git commit and I may make changes before such that fit more my coding style.

The place that I feel a bit icky around is when people do the opposite of this, where they submit a PR to some open source repo where the PR body is a huge wall of text and has stuff that normal devs don't usually write like a "files changed" section among others. Given most open source is not paid, someone picking up an issue, having Claude write the code fully, and then have Claude make the PR is disrepectful of the maintainers time and a situation I would actually call "slop", no matter if the code is resolving the issue or not. I'm okay if you use AI in the workflow, but especially if the issue is a large one, start with some discussion on how it should be resolved first rather than going in blind and putting the majority of the burden of review for the maintainer.

In summary, I want to be transparent that I am using these tools. May people call my code "slop" cause I am using these tools, maybe, but there is nuance between AI-assisted development vs fully autonomous code generation by AI models, and I want to make that apparent that I am still trying to lean towards the former.