Damus
Hoshino Lina (ζ˜ŸδΉƒγƒͺγƒŠ) 🩡 3D Yuri Wedding 2026!!! profile picture
Hoshino Lina (ζ˜ŸδΉƒγƒͺγƒŠ) 🩡 3D Yuri Wedding 2026!!!
@Hoshino Lina (ζ˜ŸδΉƒγƒͺγƒŠ) 🩡 3D Yuri Wedding 2026!!!
So I spent some time writing an AI policy for Ayagami.

TL;DR AI is incompatible with black-box/clean-room RE. If you have *ever* used AI in a Live2D/MOC3 development project, you cannot contribute, for the safety of the project.

Even if you stop using AI.

I am not hardline anti-AI; I dislike AI but this isn't a policy to screw over people who use AI. This is simply logical.

AI is trained on GitHub, including Live2D's non-free source code repositories. If you have ever read that code, you may not contribute. AI has read that code.

If this were any other project things would be different, but I am keeping things strict to avoid tainting the project with non-cleanroom code, for legal safety.

AI models are tainted, and it's transitive. Using AI models to work on MOC3 file structures taints you.

Keep in mind this ONLY applies to code/subjects relevant to MOC3/Live2D. If you ask Claude/ChatGPT/Gemini a *general* question about Rust programming, for example, that doesn't taint you (as long as you don't talk about MOC3 or Live2D or paste relevant code).
1
Hoshino Lina (ζ˜ŸδΉƒγƒͺγƒŠ) 🩡 3D Yuri Wedding 2026!!! · 3w
This is not an issue exclusive to AI. If you have ever read code related to Live2D MOC3 (e.g. another reverse engineering project), you also cannot contribute to the Ayagami core, as I cannot know that those projects themselves follow the same strict black-box policy. This strict policy applies to ...