As the organization grows into something larger than just a port of Twilight Princess, we've felt the need to create a policy we feel best represents our ideals, programming standards, and guidelines for all our future projects.
We know this is the big one. Similar to HM, we are not anti-AI. Though, we do NOT accept 100% AI-generated PRs. For menial tasks, we won't care, but submitting a PR to our repository with a Claude generated description and expecting us to give it the time of day is not what we'll ever accept. Our policy is fairly simple: if we can't tell, you can explain your change, and it's testable without us needing to do an in-depth 10 day review cycle, it's fine. We expect you to come into our project with not only the means to add your change, but the means to maintain your change. We are human-first, and slop1 will NEVER be accepted into our repositories. If you can't respect that, and must AI-generate your PRs, find another organization's project to contribute to.
We expect changes to be maintainable, architected like a reasonable SWE, and abstracted where necessary without over-abstraction. Adhere to guidelines set throughout the projects, and any individual Linifadomra project may have its own contribution guide that adds, but does not modify this policy.
Footnotes
-
AI slop for our repositories defined as code that is overly verbose, explanative, contains Claude-isms2, brittle/bug-prone, or overall reduces the maintainability of our projects. ↩ ↩2
-
A Claude-ism is a comment or markdown file that has clearly AI-generated prose, such as "not x, not y, not z", "the foo's own bar", an overuse of em-dashes or otherwise AI slop1. ↩