How to Configure JetBrains AI SSH Prevent Delete

Published
Length
3 min read

Luis Felipe Dias

All posts

To set up jetbrains ai ssh prevent delete, restrict remote host user permissions. You can also sign OpenSSH certificates that limit which shell commands run. Many engineering teams connect coding assistants to remote servers using personal keys. Those keys often map to broad Linux accounts with full write permissions. When an assistant runs a bad shell command, it can wipe critical files. How jetbrains ai ssh prevent delete works in native setups A coding tool inside an editor uses your local

AI Data Exfiltration Prevention + SSH Access Management: the complete guide

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

Free white paper. No spam. Unsubscribe anytime.

To set up jetbrains ai ssh prevent delete, restrict remote host user permissions. You can also sign OpenSSH certificates that limit which shell commands run. Many engineering teams connect coding assistants to remote servers using personal keys. Those keys often map to broad Linux accounts with full write permissions. When an assistant runs a bad shell command, it can wipe critical files.

How jetbrains ai ssh prevent delete works in native setups

A coding tool inside an editor uses your local system SSH client. It opens a remote shell and runs commands as your Linux user. The editor tool itself does not know which files are safe to remove. Instead, the remote operating system decides what that user account can do.

To prevent file removal natively, admins change permissions on the target server. First, you can create a dedicated Linux user for the coding assistant. Give this user read-only group permissions on important folders and app files. Without write permissions on a folder, the user cannot remove files inside it.

Second, teams can use OpenSSH certificates instead of static public keys. Certificate login for users and hosts arrived in OpenSSH 5.4. An admin signs a public key with a Certification Authority key. The resulting certificate holds the public key, identity info, principal names, and options. OpenSSH certificates use a much simpler format than standard X.509 certificates. Admins can embed command options inside the certificate when they sign it. When the user logs in, the server runs only the forced script.

Why host accounts and forced commands fall short

Native Linux permissions and OpenSSH certificates solve the basic login problem. They verify who logs in and which local account they can use. A certificate also lets you set a time window for that login. Still, host-level setups struggle to filter single commands inside an active session.

Developers need to write code, run tests, and restart services over SSH. If you make an account read-only, the assistant cannot help build software. If you grant write permissions for builds, the user can also delete files. Linux file permissions apply to the whole user account, not single commands.

AI Data Exfiltration Prevention + SSH Access Management: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Certificate command restrictions also create friction for interactive coding work. If you force a single script, the editor cannot run normal build tools. Admins end up building complex wrapper scripts to parse every incoming command string. Attackers or buggy prompts can bypass simple text filters using shell tricks. You also have to update that wrapper script across every remote host. Native SSH rules decide who gets in, not what each command does.

Controlling remote commands without breaking developer workflows

A safer approach separates login permissions from real-time command checks. Engineers still need an interactive session to test code and check logs. At the same time, platform teams must block destructive actions before they run. Checking each command in flight stops accidental file removal while keeping builds working. Teams also need clear logs of every command the assistant sends.

Inspect one-shot SSH commands and file tasks

SSH certificates and host rules decide who is allowed on a machine. Setup alone does not check single commands or block sftp reads on sensitive paths. By design, hoop.dev runs an open-source SSH listener on the host where people and agents keep their normal ssh, sftp, and scp tools.

First, hoop.dev guardrails check one-shot commands and file paths before they run. Next, hoop.dev data masking masks matched values in file downloads. The audit trail logs the certificate principal, requested action, and verdict without storing session content. Interactive shells carry no guardrails and can be turned off when every action must be checked. Explore the code on GitHub.

Frequently asked questions

Can JetBrains AI block file deletion on its own over SSH?

No, the assistant uses the permissions of your local SSH session. If your remote Linux user can delete files, the assistant can too. You must enforce delete blocks on the remote side or in transit.

Do OpenSSH certificates stop destructive commands by default?

No, an OpenSSH certificate only handles user or host login checks. It binds a public key to principal names and chosen certificate options. Unless you add strict command options, the user gets a normal shell.

Why do read-only Linux accounts break remote AI coding?

Coding assistants often need to edit temp files and run test suites. A strict read-only account blocks these normal tasks along with file deletion. Teams need command-level rules that allow edits but block wipe commands.

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.