Support running the agent in a remote environment (e.g., VPS)

[nezhar]

Overview

Currently, the VibePod agent runs locally. However, users may need to run the agent on a remote environment such as a VPS for a variety of practical reasons.

Motivation

There are several scenarios where running the agent remotely is preferable or necessary:

• Networking — The remote machine may have access to specific networks, VPNs, or services that are not reachable from the local machine.
• Operating System — The user's local OS may not be compatible with certain workloads. A remote Linux-based VPS can provide a consistent, supported environment.
• Access to source code or data — The source code, datasets, or files that the agent needs to work with may already reside on the remote machine, making it more practical to run the agent there directly.

Proposed Solution

• Allow vibepod-cli to target a remote environment (e.g., via SSH) when starting the agent.
• Ensure the agent can read and operate on files available on the remote host.
• Handle connectivity errors gracefully with clear, informative messages.
• Document supported remote environment configurations and setup steps.

Acceptance Criteria

• Users can point vibepod-cli to a remote host and run the agent there.
• The agent has access to the remote filesystem and environment variables as needed.
• Errors related to connectivity or misconfiguration are surfaced clearly.
• Documentation covers common remote usage scenarios and setup.