This project teaches a model how to talk to humans. The project's contributors should hold themselves to the same standard.
Be the kind of person the skill is trying to model.
- Honest. State what you actually think. Don't perform agreement.
- Specific. "This is wrong" without a reason isn't a review. Say what and why.
- Terse. Padding wastes the reader's time. So does sycophancy.
- Kind. Disagreement is fine. Cruelty isn't. The line between them is whether your comment makes the work better or makes the author smaller.
- Patient with newcomers. Everyone's first PR is messy. The second is less messy because of how the first was reviewed.
- Personal attacks, slurs, harassment.
- Demanding free labor ("just add this feature").
- Pretending to know things you don't (especially about other contributors' lives, identities, or motivations).
- Posting AI-generated comments without disclosing them (this is a project about authenticity; meta-irony aside, undisclosed AI replies on PRs are dishonest).
Maintainers will:
- Reply in-thread, naming the specific behavior and why it's not okay.
- If it continues: mute the thread, delete the comments, or remove the contributor.
- Serious or repeated violations: ban.
No three-strikes formality. Maintainer judgment, in good faith, applied case-by-case.
Email nhoxtvt@gmail.com with the subject line iamhumans CoC. Include the URL of the comment/PR/issue and a short description. We respond within a week.
Reports are kept confidential between the reporter and maintainers unless the reporter explicitly authorizes wider disclosure.
This applies to all repo surfaces: PRs, issues, Discussions, commit messages, code comments, README, and any other channel the project owns. It does not apply to private conversations between contributors elsewhere.
Influenced by the Contributor Covenant, but kept short on purpose. A long code of conduct that nobody reads protects nobody.