CODEOWNERS Manifesto

A clear stance on code ownership: if you build it, you own it—covering responsibility, quality, and long-term accountability.

1. Code Ownership

   Each part of the codebase must have a clearly designated owner. This ownership extends to all outputs influencing the system, including AI-generated code. Accountability cannot be delegated to tools.

   Any team member may contribute to any part of the codebase, but all changes must be reviewed and approved by the responsible code owner.

   Contribution is open; accountability is not.

2. Code Quality

   Code owners are responsible for ensuring the quality of the code they own. This includes ensuring it meets coding standards, passes all tests, and is well-documented, whether the code is human-written or AI-generated/AI-assisted.

3. AI Usage Responsibility

   • Contributors must use AI responsibly, understand the limitations and risks (hallucinations, bias, outdated patterns) and verify outputs

     Both contributors and code owners share responsibility for AI usage:

   • Code owners must critically evaluate AI-assisted contributions

   AI accelerates contribution—but does not reduce the need for judgment.

4. Review Process

   All changes must be reviewed by the code owner, including AI-assisted contributions. This ensures that only changes that meet quality standards and align with the project's direction are merged. Reviews should explicitly consider:

   • Whether AI-generated code is correct and necessary

   • Hidden complexity or subtle errors introduced by AI

   • Long-term maintainability (not just short-term productivity)

5. Transparency

   Use of AI in code contributions should be transparent. Contributors should disclose when AI tools were used so reviewers can apply appropriate scrutiny during reviews.

This principle will be enforced. The implementation approach is under development and will be published shortly.

1. Collaboration

   Code owners should actively guide contributors—especially in a shared environment:

   • Provide clear expectations and standards

   • Help contributors navigate the codebase

   • Encourage good use of AI, and challenge poor use

2. Learning and Improvement

   Both contributors and code owners are expected to continuously improve:

   • Understanding of the codebase

   • Effective and safe use of AI tools

   • Ability to review and write high-quality code

   • Skill in evaluating AI-generated output

   AI is a multiplier—but only for those who know how to evaluate it.

3. Respect and Communication

   Discussions about code—including AI-generated contributions—should remain respectful and constructive. Questioning AI output is expected and encouraged.

4. Responsibility

   Being a code owner means owning outcomes, not just inputs.

   A shared codebase requires strong collaboration.

   Feedback should be direct, constructive, and respectful. Challenging code—including AI-generated code—is part of the process.

5. Adaptability

   Code owners must adapt to evolving tools, including AI-assisted development. Teams should continuously refine how ownership, contribution, and AI usage work together.