Feedback requested: Anti AI Spam Policy

[nemesifier]

I opened openwisp/openwisp-docs#286 to add an Anti AI Spam Policy to the OpenWISP documentation.

The goal is to protect maintainers and project infrastructure from low-effort, careless, or mostly AI-generated contributions that shift the burden of understanding, implementation, testing, and review onto the project.

The proposed policy states that OpenWISP welcomes meaningful contributions from people who understand the project, the issue, and the changes they submit. It also makes clear that AI tools do not replace software engineering experience or responsibility: contributors must be able to explain, maintain, and test the code they propose.

The policy gives maintainers explicit grounds to close pull requests considered spammy or harmful, including:

• unsolicited PRs without a related issue;
• PRs targeting issues which have not been opened, acknowledged, validated, or accepted by maintainers;
• poor, incomplete, or untested implementations;
• PRs generated by AI tools or automation without adequate human understanding, review, and testing;
• repeated pushes, repeated failed CI builds, or excessive automated review activity which wastes maintainer time and consumes limited sponsored CI and review tool credits;
• PRs which show little or no evidence that the contributor understands the project, the issue, or the proposed solution.

The policy also states that maintainers may temporarily limit contributions from contributors who violate it, and may permanently ban contributors who continue violating it after a temporary restriction. When a PR is closed under this policy, maintainers may provide only a link to the policy as the explanation.

Feedback requested:

• Is the policy clear enough?
• Is it strict enough to protect maintainers and sponsored infrastructure?
• Is any wording likely to be misunderstood or abused?
• Are there important cases we missed?

Please leave any objections, concerns, or suggested improvements in the PR or in this discussion. If no substantive objections are raised after a 2 weeks feedback window, I will treat that as no objection from the community and proceed with the addition of the new policy.

[okraits]

For me it's fine. Maybe it would be beneficial to emphasize that things like the format/qa scripts and the tests must be run locally before pushing the code to keep load from the CI.

And I was just thinking of additional cases which might come with the use of AI:

• pull requests which don't align with the general coding style of the project
• pull requests which change or add more than what would be needed for the problem to solve

Regarding the general topic of AI spam, too little understanding of the project and the submitted changes and the burden on the maintainers, I saw two examples in the recent past:

• the curl project takes a month off from processing vulnerability reports: https://daniel.haxx.se/blog/2026/06/15/curl-summer-of-bliss/
• the Ladybird browser project will no longer accept public pull requests at all: https://ladybird.org/posts/changing-how-we-develop-ladybird/

[nemesifier]

Published https://openwisp.io/docs/dev/general/code-of-conduct.html.