Visibility into the future of the codex-rs harness

What feature would you like to see?

Now that OpenAI supports 3rd party OSS harnesses, it would be good for the firm to communicate the bare minimum possible about what the aims are regarding this harness. It has different features than the 3rd party OSS ones, and users might find the 3rd party ones more attractive for some use cases. Migration is easy but not costless (eg. other harnesses support additional frontmatter fields for slash commands / prompts).

Will codex-rs attempt to have feature parity with other harnesses (and generally be more feature rich than it currently is)?
Or will it be more of a barebones reference implementation?
Will codex-rs have differentiating features, such as early access to newer models from OpenAI, better compaction, or other API-provided services? I understand that disclosing specific information in such a dynamic and competitive market is unfeasible - I'm just asking for the minimum information to decide on whether to continue using codex-rs [*] or try other harnesses.

Thank you.
[*] codex-rs has been my exclusive agent since the launch of gpt-5, and I've accomplished a great deal thanks to it - so thank you and congratulations for all the work on this project.

Additional information

_No response_