fix(ops): fix AGPL-3.0 recognition, add AI training clause and .aiignore

[forkwright]

Problem

Three issues with the current license setup:

1. GitHub can't parse the LICENSE file — it shows as "other" instead of AGPL-3.0. The file uses a custom copyright header format instead of the FSF's standard AGPL-3.0 full text. This means GitHub's license detection fails, which affects how the repo is displayed and indexed.

2. No AI training protection — AGPL-3.0 does not restrict use of the code as machine learning training data. The codebase (RF intelligence, mesh networking, communications sovereignty) is exposed to AI training crawlers with no legal restriction.

3. No .aiignore — GitHub's AI pipelines (Copilot, Models) can ingest repo content. The April 24, 2026 policy change makes this opt-out, not opt-in.

Solution

Fix LICENSE file

Replace the custom-header LICENSE with the standard FSF AGPL-3.0 full text so GitHub's license detection recognizes it. Add a copyright notice at the top in the standard format:

Copyright (C) 2026 Cody Kickertz

This program is free software...

Add AI training prohibition clause

Append as a supplemental notice (not modifying the AGPL text itself, which the FSF prohibits):

Create a NOTICE file or add to README:

Supplemental to the AGPL-3.0 license: Use of this work or its contents, in whole or in part, to train, fine-tune, distill, or otherwise improve machine learning or artificial intelligence models is expressly prohibited without prior written consent from the copyright holder.

NOTE: The AGPL text itself cannot be modified per FSF terms. The AI restriction is a supplemental condition from the copyright holder, not a license modification.

Add .aiignore

*

Impact

• GitHub correctly identifies the license (visibility, compliance tooling, search)
• AI training prohibition creates enforceable terms (EU) and cause of action (US)
• .aiignore blocks GitHub's own AI pipelines

Relates to

• forkwright/kanon#426 — kanon license structure