sbx-agents 0.1.0

Python SDK for running coding agents in sandbox backends

sbx-agents

Build with Codex, OpenCode, Claude Code, or your favorite AI coding agent from Python.

sbx-agents runs coding agents in isolated sandboxes and gives your app structured access to output, logs, diffs, files, and events.

Install

With uv:

uv add sbx-agents

With pip:

pip install sbx-agents

Install the sbxa CLI as a standalone uv tool:

uv tool install sbx-agents
sbxa --help

For local development:

uv tool install -e .
sbxa --help

Basic Example

from pathlib import Path

from sbx_agents import Agent, Runner
from sbx_agents.backends import Codex
from sbx_agents.sandboxes import DockerSbxSandbox

agent = Agent(
    name="fix-tests",
    backend=Codex(model="gpt-5-codex"),
    instructions="Fix failing tests with the smallest safe diff.",
)

result = Runner.run_sync(
    agent,
    sandbox=DockerSbxSandbox(
        workspace=Path("."),
        branch="agent/fix-tests",
    ),
    prompt="Fix the failing pytest suite.",
)

print(result.output)
print(result.diff)

Define a reusable runner and launch it later with sbxa:

runner = Runner(
    agent,
    sandbox=DockerSbxSandbox(workspace=Path("."), branch="agent/fix-tests"),
    default_prompt="Fix the failing pytest suite.",
)

sbxa run main:runner

Backends

Agent backends:

• Codex
• OpenCode
• Shell
• custom backends

Sandbox backends:

• DockerSbxSandbox
• DockerSandbox
• custom backends

Docs

Full docs, examples, API reference:

https://blairhudson.com/sbx-agents/