How to Configure Devin SSH Connect

Published
Length
3 min read

Luis Felipe Dias

All posts

You set up a devin ssh connect workflow by giving the agent an SSH key or certificate. Then you allow its IP address to reach your target server or bastion host. Engineering teams often start by adding a static public key to a server user account. This fast setup lets the autonomous agent log in and run shell commands right away. Still, broad server roles create big risks when an automated tool can run any command. How a native devin ssh connect setup works In a standard setup, the AI agent r

SSH Access Management + End-to-End Encryption: the complete guide

Architecture patterns, implementation strategies, and security best practices. Delivered to your inbox.

Free white paper. No spam. Unsubscribe anytime.

You set up a devin ssh connect workflow by giving the agent an SSH key or certificate. Then you allow its IP address to reach your target server or bastion host. Engineering teams often start by adding a static public key to a server user account. This fast setup lets the autonomous agent log in and run shell commands right away. Still, broad server roles create big risks when an automated tool can run any command.

How a native devin ssh connect setup works

In a standard setup, the AI agent runs inside its own workspace with an SSH client. The agent needs a private key to log in to your remote Linux hosts. You place the matching public key on the target server under a specific Linux user. When the agent starts a task, it opens an SSH tunnel to the server.

Many teams prefer OpenSSH certificates over static keys for better security. In OpenSSH, a Certification Authority key signs a public key to create a certificate. This certificate holds the public key, identity information, principal names, and a set of options. OpenSSH certificates use a much simpler format than X.509 certificates. You can sign a user certificate and list one or more allowed user principals. The server trusts the authority key instead of tracking many single public keys.

You can also place a jump host between the agent and your private network. The agent first logs in to the bastion and then hops to the final host. This network path keeps your internal servers off the public internet. System admins set firewall rules so only the agent workspace can reach the entry host.

What native SSH controls solve and what they miss

Native OpenSSH features do a great job of checking who can log in. Keys and certificates prove the identity of the agent before a shell opens. Linux user accounts and group rules limit which files the agent can read or change. If you use OpenSSH certificates, you can set a time window so access expires.

Still, login checks do not control what individual commands do after the session starts. Once the agent has an active shell, it holds the full power of that user. An autonomous coding agent runs fast and executes many shell commands in seconds. A bad prompt or logic loop might tell the agent to drop a local file. It could also read sensitive config files or restart a busy production service.

SSH Access Management + End-to-End Encryption: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Standard server logs also fall short when you review automated sessions. Basic system logs show when a user logged in and when the session ended. They do not always give a clear view of every command and its output. If an agent queries a local tool, raw data like emails or phone numbers returns. Card numbers or national IDs can also flow right back into the agent context. Native SSH setups cannot rewrite matched fields in results or pause risky commands in flight.

Securing agent shell sessions in practice

To reduce risk, teams must separate login rights from runtime command control. Start by giving the agent a dedicated Linux user with minimal file permissions. Never give an autonomous agent full root rights on a production machine. Use OpenSSH certificates with short validity windows instead of permanent static keys. Keep the agent out of production networks unless a task requires it. Even with these steps, teams still need visibility into every command the agent sends.

Add command and file guardrails to SSH access

Host firewalls and certificate principals control who opens an SSH session. Yet an allowed user or agent can still run a bad one-shot command or pull sensitive files over sftp. In practice, hoop.dev provides an open-source SSH server with guardrails on the host that uses OpenSSH certificates.

Using hoop.dev SSH controls, guardrails check one-shot commands, env variable names, and sftp or scp paths before they run. Port forwards default to deny unless the target is listed. Because interactive shells have no statement boundary and carry no guardrails, teams can turn them off so every command is checked and logged. Check out the MIT-licensed project on GitHub.

Frequently asked questions

How does Devin log in to a private server over SSH?

The agent uses an SSH private key stored in its workspace to log in. Your server checks the matching public key or OpenSSH certificate before opening a shell.

Why are static SSH keys risky for autonomous AI agents?

Static keys do not expire on their own and grant full user shell access. If an agent runs a bad command, the server cannot block that single action.

Can OpenSSH certificates limit which user accounts an agent can use?

Yes, an OpenSSH user certificate lists allowed principal names signed by your authority key. The target server only lets the agent log in as those specific principals.

Save the open-source gateway.

hoop is MIT-licensed infrastructure for controlling how AI agents reach production data. Star hoophq/hoop so you can inspect it, deploy it, or share it when your team starts governing agent access.