Thanks for contributing to SWE-AF.
- Python 3.12+
- AgentField control plane (
af) - Access to an AI runtime used by your run (
claude_codeoropen_code)
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -e ".[dev]"- Start AgentField control plane:
af- In a second terminal, start this node:
python3 -m swe_af- Run tests before opening a PR:
python3 -m unittest discover -s tests -v- Do not commit generated Python caches or Rust
target/outputs. - Keep example artifacts that document completed example runs.
- Keep changes scoped; avoid unrelated formatting churn.
Each PR should include:
- What changed and why.
- Any behavior changes to agent orchestration.
- Test evidence (unit tests and/or command output).