Thank you for contributing to a cntryl project. We welcome thoughtful contributions from people using any development workflow, including AI-assisted workflows. AI assistance is not a negative signal. We require transparency, useful context, and contributor accountability for every submission.
Repository-specific instructions take precedence for project setup, validation, architecture, and style. The expectations in this document apply throughout the cntryl organization.
Explain why you are making the contribution:
- How did you encounter the problem or need?
- What project, workflow, or use case does it affect?
- What outcome do you expect?
- Why is the proposed scope appropriate for this repository?
Provide a minimal reproduction, concrete example, measurements, or other supporting evidence when applicable. You do not need to disclose your legal identity, private information, employer, or customer information.
Open an issue and wait for maintainer agreement before beginning a substantial feature, architectural change, broad refactor, compatibility change, or other large body of work. A small, self-contained correction may be submitted directly when its motivation and evidence fit clearly in the pull request.
Disclose whether an AI or other generative tool materially assisted the issue, pull request, code, documentation, tests, design, or analysis. When it did, briefly describe:
- what kind of tool was used;
- which parts of the contribution it assisted; and
- how you reviewed and validated the resulting work.
Material assistance includes generating or substantially rewriting submitted content or supplying analysis on which the contribution relies. Routine editor completion, formatting, spelling correction, and search generally do not need individual disclosure.
Do not include private prompts, credentials, proprietary source material, or other confidential information in the disclosure.
By submitting a contribution, you are expected to:
- understand the complete change and explain its important behavior and design;
- review the full diff rather than forwarding generated output unexamined;
- run appropriate tests and report the exact validation performed;
- distinguish observed results from assumptions or unverified claims;
- respond to maintainer questions and make reasonable revisions; and
- have the right to submit the work under the repository's license.
The contributor remains responsible for the submission regardless of which tools assisted it.
Search existing issues and pull requests before opening a new one. Use the provided form or template and complete every applicable section.
Keep pull requests focused. Link the accepted issue for substantial work, describe what changed and why, identify compatibility or security risks, add or update tests, and list the commands used for validation. Do not claim that a test or check passed unless you ran it against the submitted change.
Every contribution is evaluated on its technical merit, supporting evidence, scope, contributor engagement, and project fit. AI disclosure does not count against a contribution. Submission volume, prior trivial changes, polished prose, account characteristics, or apparent authorship do not create an entitlement to trust or merge.
Maintainers may close submissions that are duplicate, speculative, misleading, mass-produced, unrelated to a demonstrated need, too broad without prior agreement, missing required context or disclosure, or unsupported by a contributor who cannot explain or revise the work. This protects limited review capacity and the integrity of the project; it is not a judgment about a contributor's identity or choice of tools.
Maintainers make the final decision about whether and when a contribution is accepted.