WorldForge is a technical project for physical-AI and world-model tooling. Discussion should stay professional, evidence-based, and welcoming to contributors with different backgrounds.
- Be direct without being hostile.
- Critique code, claims, designs, and experiments rather than people.
- Keep public issues and pull requests focused on reproducible facts, tradeoffs, and next steps.
- Respect that optional robotics and model runtimes can be expensive, fragile, and environment-specific; ask for clear evidence before assuming user error.
- Disclose conflicts of interest when promoting an adapter, benchmark, or provider result.
- Harassment, threats, insults, or discriminatory language.
- Publishing private credentials, checkpoints, datasets, security reports, or personal data.
- Repeatedly derailing issue or pull request threads after maintainers have set scope.
- Presenting scaffold adapters, deterministic demos, or synthetic benchmark results as real upstream model performance.
Maintainers may edit or hide comments, lock threads, close issues, reject contributions, or block participants when behavior undermines the project. Security reports should follow SECURITY.md, not public issues.