How kanman marks its work

Everything kanman writes into your repositories and your tracker is marked as AI work, so reviewers, auditors and your own tools can always tell it apart from work people did. The marking follows the EU AI Act’s transparency rules for AI-generated content and works the same on GitHub and GitLab, on kanman Cloud and on your own runners.

Commits

Every commit kanman pushes ends with a trailer that names kanman, the coding agent and the model it used:

Assisted-by: kanman (claude_code/claude-sonnet-4-5)

The trailer is in the form many open source projects already use for AI assistance. kanman never adds a Signed-off-by line: certifying the Developer Certificate of Origin is something only a person can do, so a sign-off in a commit kanman wrote is removed before the push. The commit author stays kanman’s own identity.

Search for kanman’s commits with git log --grep "Assisted-by: kanman".

Pull requests and merge requests

Every pull request or merge request kanman opens, including reverts:

• carries the label kanman, so you can filter for it in your code host and in branch protection or CI rules;
• ends with one line in its description: AI-generated: kanman, an AI agent, prepared this.

If your repository does not allow kanman to set labels, the pull request still opens and the line in the description still marks it.

Checked on every change

kanman does not only add the marking, it checks it where it lands. When it reviews a change, it reads what your code host shows: whether each of its commits carries the trailer, and whether the pull or merge request carries the label and the line. Every week the team’s report counts the changes kanman merged and how many of them were fully marked, and kanman’s operators are alerted about any change that was not.

Tracker comments

Comments kanman writes on your Jira, GitHub or GitLab issues end with the same line.

What is not marked

The marking describes kanman’s work, never a person. kanman does not rate, rank or score the people on your team, and its reports describe how a team delivers, not how individuals perform.