Background
airgap-coder enables Codex CLI to run in restricted or air-gapped environments through LiteLLM and self-hosted OpenAI-compatible model services such as vLLM and SGLang.
To validate the project in real environments and build a sustainable maintenance process, we are looking for 3–5 long-term contributors or maintainers. You are welcome to help with environment testing, code, documentation, issue triage, PR review, or releases.
Areas where help is needed
- Validate the Codex CLI → LiteLLM → vLLM/SGLang compatibility path;
- Test core workflows such as
lc init, doctor, test, and export;
- Maintain a compatibility matrix across Codex, LiteLLM, models, and operating systems;
- Improve offline packaging, credential isolation, automated testing, and security boundaries;
- Triage issues, review PRs, improve English/Chinese documentation, and prepare releases.
Collaboration and permissions
Collaboration can begin through issues and pull requests. Based on sustained contributions and mutual agreement, contributors may gradually take on project maintenance responsibilities with appropriate GitHub Triage or Maintain permissions.
Contributions will be credited in the project documentation and release notes.
Privacy and security
Contributors never need to disclose real credentials, internal network addresses, or private source code. All logs, configurations, and environment details must be sanitized. Testing can also be performed entirely with project-provided mock environments.
OpenAI Codex for Open Source
The project has applied to OpenAI Codex for Open Source.
If you later become a genuine core maintainer, contribute consistently, and hold the appropriate repository permissions, you may submit your own application using your ChatGPT account and OpenAI Organization ID, or be nominated by an existing maintainer. Applications must accurately describe your role, maintenance responsibilities, and repository access.
OpenAI reviews each maintainer independently. Any approved ChatGPT Pro, Codex, Codex Security, or API credit benefits are personal and non-transferable; they cannot be shared, transferred, or pooled.
How to join
If you are interested, please reply to this issue with:
- The platforms or model backends you can test;
- Whether you prefer testing, coding, documentation, or maintenance work;
- Roughly how often you may be able to participate.
We would be happy to have you help make airgap-coder a reusable, verifiable, and sustainably maintained Codex tool for restricted environments.
Background
airgap-coder enables Codex CLI to run in restricted or air-gapped environments through LiteLLM and self-hosted OpenAI-compatible model services such as vLLM and SGLang.
To validate the project in real environments and build a sustainable maintenance process, we are looking for 3–5 long-term contributors or maintainers. You are welcome to help with environment testing, code, documentation, issue triage, PR review, or releases.
Areas where help is needed
lc init,doctor,test, andexport;Collaboration and permissions
Collaboration can begin through issues and pull requests. Based on sustained contributions and mutual agreement, contributors may gradually take on project maintenance responsibilities with appropriate GitHub Triage or Maintain permissions.
Contributions will be credited in the project documentation and release notes.
Privacy and security
Contributors never need to disclose real credentials, internal network addresses, or private source code. All logs, configurations, and environment details must be sanitized. Testing can also be performed entirely with project-provided mock environments.
OpenAI Codex for Open Source
The project has applied to OpenAI Codex for Open Source.
If you later become a genuine core maintainer, contribute consistently, and hold the appropriate repository permissions, you may submit your own application using your ChatGPT account and OpenAI Organization ID, or be nominated by an existing maintainer. Applications must accurately describe your role, maintenance responsibilities, and repository access.
OpenAI reviews each maintainer independently. Any approved ChatGPT Pro, Codex, Codex Security, or API credit benefits are personal and non-transferable; they cannot be shared, transferred, or pooled.
How to join
If you are interested, please reply to this issue with:
We would be happy to have you help make airgap-coder a reusable, verifiable, and sustainably maintained Codex tool for restricted environments.