Use this before publishing the repository.
- Confirm repository name and description.
- Review README positioning.
- Confirm
LICENSEis correct. - Review
CONTRIBUTING.mdclean-room rules. - Review
SECURITY.md. - Run the full test suite.
- Run mock CLI smoke.
- Run mock eval smoke.
- Run DeepSeek Lab mock smoke.
- Confirm no API keys or secrets are present.
- Confirm
runs/andeval-runs/are ignored. - Confirm GitHub Actions CI passes after push.
- Confirm issue and PR templates render on GitHub.
- Review
docs/release.md.
- Add GitHub Actions CI.
- Add issue templates.
- Add pull request template.
- Add a first release tag.
- Add example provider configs.
- Add a project roadmap.
$env:PYTHONPATH='src'
python -m unittest discover -s tests
python -m opencode_harness version
python -m opencode_harness eval examples/mock-suite.json --preset mock --max-steps 2 --context-chars 1000
python -m opencode_harness eval model-labs/deepseek/mock-smoke-suite.json --preset mock --max-steps 2 --context-chars 1000
python -m opencode_harness replay runs/latest.jsonl --summary
python -m opencode_harness dashboard eval-runs --output eval-runs/dashboard.htmlRecommended short description:
A clean-room, model-agnostic harness for Claude Code-class coding agents, with DeepSeek/Qwen/Claude/OpenAI backends, MCP tools, trace replay, and evals.
Recommended disclaimer:
This project is a clean-room implementation. It does not contain, derive from, or depend on leaked or proprietary Claude Code source code.