This document is heavily inspired by the Linux Kernel and LLVM guidelines.
Eter empowers contributors to use the tools of their choice, provided that a human-in-the-loop remains the fundamental requirement. Every piece of code or text generated by an LLM must be thoroughly scrutinized and validated by the contributor before it is submitted for peer review.
As a contributor, you are the sole author and are fully accountable for your work. Before requesting a review, ensure your submission meets our quality standards. Maintainer time is a finite resource; you must be prepared to defend and explain your technical choices during the review process.
- For New Contributors: We encourage starting with small, manageable tasks. Mastery comes from understanding every line you write. Using an AI to bypass the learning process—or simply funneling maintainer feedback back into an LLM—stalls your professional growth and undermines the health of our community.
- Mentorship: We strive to be a welcoming project. We invest in newcomers to help them become future maintainers, but this relationship relies on the contributor’s active effort to learn and iterate based on feedback.
Contributors are required to disclose and label any submissions containing significant AI-generated content. This transparency is not for tracking purposes, but to provide reviewers with the necessary context to evaluate the work effectively.
Please note the use of AI tools in your Pull Request description or commit messages using the following trailer format:
Assisted-by: AGENT_NAME:MODEL_VERSION [ADDITIONAL_TOOLS]
Example:
Assisted-by: Claude:4.7-Sonnet
Assisted-by: GitHub-Copilot:auto
The Signed-off-by and Co-authored-by trailers are reserved for human contributors and should not be used to indicate AI assistance.
Our guiding principle is simple: A contribution must provide more value to the project than the effort required to review it.
This idea is captured by this quote from the book Working in Public by Nadia Eghbal:
"When attention is being appropriated, producers need to weigh the costs and benefits of the transaction. To assess whether the appropriation of attention is net-positive, it’s useful to distinguish between extractive and non-extractive contributions. Extractive contributions are those where the marginal cost of reviewing and merging that contribution is greater than the marginal benefit to the project’s producers. In the case of a code contribution, it might be a pull request that’s too complex or unwieldy to review, given the potential upside." – Nadia Eghbal
We define "extractive contributions" as unverified AI outputs that shift the burden of work from the author to the maintainer. While AI can accelerate development, it must not be used to flood the project with low-effort PRs or RFCs. To protect our maintainers from burnout, we prioritize quality and human oversight over sheer volume.
The legal landscape regarding AI and copyright is complex. Eter’s policy is strictly aligned with our licensing standards:
- Ownership: You must ensure you have the legal right to contribute the code (whether through personal ownership, employer permission, or open-source compatibility).
- Infringement: Using an AI to reproduce copyrighted material does not strip that material of its original protections. Contributors are responsible for ensuring no infringing content enters the codebase.
- Remediation: Any contribution found to violate copyright or these guidelines will be removed immediately.