Trial.run() call, and a task is fully described by four files
(instruction.md, Dockerfile, test.sh, task.toml), so mixed task suites —
SWE-bench, Terminal-Bench, custom tasks — train through one endpoint.
Miles integrates Harbor as an
agent-function integration: the agent function
hands each session’s OpenAI-compatible URL to Harbor, which runs the per-task
sandbox, runs the agent against that URL, and grades the result; the grade
becomes the sample’s reward through a custom reward hook.
Try it
Two execution modes; each example README is the complete guide for its mode:- Tasks on a local Docker daemon → the agent-server mode:
examples/swe-agent-harbor-docker. - Tasks on a cloud sandbox provider → the
in-process mode, no server in between:
examples/experimental/harbor.