Some teams ask for a note in the pull request whenever an AI agent wrote part of it. Some add a co-author line to commits automatically. Others treat it like any other tool: nobody labels code that came from autocomplete or a snippet library, so why label this?
The question matters more for reviewers than for credit. A reviewer who knows a block came from an agent might read it differently: check that the imported function exists, look for tests that only test the mock, ask whether the author understands the change. A label can also turn into noise if it ends up on every pull request, or into a stigma if it makes people hide it.
We would like to hear what your team does:
- Do you mark agent-written code, and where: the pull request description, commit trailers, nowhere?
- Has knowing a change came from an agent ever changed how you reviewed it?
- If your team added or dropped such a rule, what pushed the change?
Please share your team's practice in the comments below, including the arguments that settled it.


