AI Tool Usage

Usage policy How AI tools may be used in this project

AI-generated code Acceptance policy for AI-authored contributions

AI in CI/CD Can AI tools run in automated pipelines

Training data opt-out Request that project code not be used for AI model training

AI disclosure Assisted-by commit trailer policy for AI-assisted work

Agent Rules

Code style enforcement Style rules agents must follow (leave empty for none)

Test requirements Testing expectations for AI-generated changes

Restricted paths Paths agents should not modify (comma-separated)

Review requirements Human review policy for AI-generated pull requests

CLAUDE.md Handling

Generate as Separate file or symlink instruction pointing to AGENTS.md

What is AI_POLICY.md?

A convention for repositories to declare how AI tools interact with their codebase -- which tools are permitted, how AI-generated code is handled, and training data preferences.

Who's adopting this?

Projects like CloudNativePG, Kyverno, and Kubewarden ship AI policy files. Templates here track current patterns.

Why add AI policy files?

Sets expectations for contributors and gives maintainers a governance reference point.

Convention references

Informed by the AGENTS.md spec, Linux Foundation AI governance guidance, and CNCF project practices. See conventions.json for the full source list.