MagicAI welcomes contributors, players, judges, developers, researchers, and documentation writers. Participation must remain respectful, constructive, and safe.
Expected behavior includes:
- discussing ideas and code without attacking people;
- giving specific, actionable feedback;
- respecting different experience levels, languages, identities, and play communities;
- acknowledging uncertainty and correcting mistakes openly;
- protecting private information and unpublished vulnerability details;
- crediting the work and sources of others.
Unacceptable behavior includes harassment, discrimination, threats, deliberate misinformation, sexualized conduct, personal attacks, doxxing, spam, or attempts to pressure maintainers into unsafe releases.
Rules disagreements must be resolved through evidence and current authoritative sources, not status or popularity. Strategic disagreement is welcome when it remains clearly separated from factual authority.
Evaluation artifacts, community examples, and reported failures must not be repurposed as training data without explicit permission and a documented policy change.
Maintainers may edit, hide, lock, or remove contributions and may temporarily or permanently restrict participation when behavior violates this policy.
Report conduct concerns privately through the same channel described in SECURITY.md, while clearly identifying the report as a conduct matter rather than a software vulnerability.