Contribute to the lab

New contributors can begin with documentation, offline tests or research proposals. AI coding assistants are optional, and meaningful technical verification remains a human responsibility.

Find a small, reviewable task

Check good first issues, help wanted and the contribution guide. Read the issue acceptance criteria before beginning; never infer the task from its title alone.

Optional Claude Code / Codex workflow

Both tools can help inspect code, propose test cases and draft docs, but AI-produced code and measurements require human checking. The repository's AI-assisted contribution tutorial includes safe examples, while agent prompts provide ready-to-adapt templates.

# Example prompt for a coding assistant:
Read AGENTS.md and issue #NUMBER.
Inspect only the relevant files.
Propose the smallest change and a verification plan.
Do not claim cross-worker DDS delivery without an integration test.
Never expose or commit Kubernetes credentials.
Describe what was tested and what remains unverified.

Quality checklist

• All commands reference real checked-in filenames and paths.
• Version and support claims are linked to verified evidence.
• ROS 2 talker/listener communication is never implied by running stress-ng.
• Keep credentials, join tokens, kubeconfigs and user data out of public issues and logs.
• Preserve Apache-2.0 licensing and attribution for external assets.
• Honor reduced-motion, mobile layout and accessible navigation for website changes.
• A Docker build or YAML lint does not establish full cluster compatibility.

Research, teaching and collaboration

University courses, ROS middleware researchers and infrastructure teams can share concrete requirements through GitHub issues. Contributions and discussions are public, so share only non-sensitive information. This website does not advertise a paid service, commercial support agreement or active product offering.