Execute Commands Inside Your Sandbox

Execute short, non-streaming commands within a Fermion Sandbox and return stdout and stderr once complete using the Python SDK.

The run_command() method executes a short-lived command inside your active sandbox container and waits for it to finish before returning the output.

It’s ideal for quick, synchronous tasks such as listing files, checking Python or Node.js versions, or running lightweight scripts that complete within a few seconds.

This method captures both stdout and stderr once the command completes. Because it’s optimized for speed and simplicity, run_command() has a strict 5-second execution limit and does not support streaming output.

---

async run_command(cmd: str, args: Optional[List[str]] = None) -> CommandResult

Description:
Executes a short command (< 5 seconds) and returns its standard output and standard error.
For longer-running or streaming processes, use run_streaming_command() instead.

---

Parameters:

Parameter	Type	Required	Description
cmd	str	Yes	Command to execute (e.g. "python3", "ls", "echo")
args	List[str]	No	List of command arguments

---

Returns:

CommandResult object with the following attributes:

• stdout: str - Standard output
• stderr: str - Standard error
• exit_code: int - Exit code (defaults to 0)

Example:

import asyncio
from fermion_sandbox import Sandbox

async def main():
    sandbox = Sandbox(api_key="your-api-key")
    await sandbox.create(should_backup_filesystem=False)

    # Run a simple shell command
    result = await sandbox.run_command(cmd="pwd")

    print("Current directory:", result.stdout)
    print("Exit code:", result.exit_code)

asyncio.run(main())

Limitations:

• 5 second maximum : Command will timeout
• No streaming - must wait for completion
• Exit code is always 0 for this method (use run_streaming_command() if you need accurate exit codes)