How to Configure Windsurf SSH Audit Log

Published
Length
3 min read

Luis Felipe Dias

All posts

You set up a windsurf ssh audit log by turning on verbose system logs on your target Linux servers. You also need to sign user keys with a central authority so each login records a clear user name. Many teams let the AI agent use a developer's local shell keys to reach remote hosts. When those keys have broad rights, the agent can run any shell command on the server without limits. How a native windsurf ssh audit log works Cognition renamed Windsurf to Devin Desktop in June 2026, while keepin

Audit Log Integrity + 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.

You set up a windsurf ssh audit log by turning on verbose system logs on your target Linux servers. You also need to sign user keys with a central authority so each login records a clear user name. Many teams let the AI agent use a developer's local shell keys to reach remote hosts. When those keys have broad rights, the agent can run any shell command on the server without limits.

How a native windsurf ssh audit log works

Cognition renamed Windsurf to Devin Desktop in June 2026, while keeping the same IDE features. Inside the app, the agent can run terminal commands on its own after the user gives permission. It uses a dedicated shell session that loads the user's normal shell profile, aliases, and variables. Because the agent shares the user's shell, it also uses the user's local SSH keys and config.

When the agent connects to a remote machine, the target server's SSH daemon handles the login check. By default, standard static public keys only tell the server which key matched a local server account. If three people share one server account, the server log only shows that shared account name. To fix this, teams use OpenSSH certificates instead of plain public keys.

An OpenSSH certificate holds a public key, an identity string, allowed user names, and signing options. A central authority key signs the user key and embeds the developer's specific name inside it. When the agent connects using this signed key, the remote server writes that exact identity into the system log. The admin can then forward those system logs to a central log collector for storage and review.

What native SSH logging solves and what it misses

Native OpenSSH logging and signed certificates solve the problem of tracking who logged into a host. You can check the server log and see when a user's key opened a session. You can also see the source address, the key identity string, and when the connection closed. This setup stops anonymous logins and helps teams retire shared static keys on their servers.

Still, the native SSH daemon only logs the login event and the session end time. Once the shell opens, the SSH daemon creates a tunnel and stops reading the traffic inside it. It does not record the specific bash commands that the AI agent runs during that session. If the agent deletes a folder or reads a secret file, the SSH log will not show it.

Audit Log Integrity + SSH Access Management: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Also, native server logs cannot tell a human typing apart from an AI agent running commands. Both the developer and the agent use the exact same local terminal setup and the same signed key. Even if you add host-level process tools, those tools reside on the target server itself. A user or an agent with root rights on that host can turn those local log tools off.

Why agent shell commands need runtime controls

AI coding agents work fast and can chain many shell commands together in a few seconds. A user might approve a local command without seeing that it opens a remote shell to production. Once connected, a standard SSH role gives the agent the full rights of that remote Linux user. There is no built-in way for OpenSSH to check each command inside an open interactive shell.

To keep servers safe, teams need a way to inspect and log every command before it runs. They also need a way to mask sensitive data like emails, phone numbers, and card numbers in command output. Putting these checks outside the target server ensures that even a root user cannot erase the audit trail.

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 inline open-source SSH listener on the host where engineers and AI agents keep their normal ssh, sftp, and scp tools. On every production command or file operation, hoop.dev decides whether to pass, mask, block, or hold it for approval.

First, hoop.dev guardrails check one-shot commands and file paths before they run, and 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

Does the Windsurf IDE keep its own remote SSH log?

No, the IDE runs terminal commands inside your local shell and does not record remote server activity. You must capture the session logs on the remote host or in your network path.

Can OpenSSH log every command an AI agent runs?

No, the OpenSSH server daemon only records login events, key details, and session start or end times. It does not inspect or log the individual shell commands sent inside an interactive terminal session.

You can sign each user's public key with an OpenSSH certificate authority and set a unique identity string. The remote server will write that identity string and the key fingerprint into its login log.

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.