AI / Agents Agents

Cite as: Real Problem AI problem “How do I stop my AI agent from using my SSH keys on remote machines?”. Opportunity score 7.8 out of 10 (severity 8, AI feasibility 8, market signal 8, competition gap 7). Category AI / Agents. Trend Agents. Source signal: Hacker News thread "ChatGPT Desktop forwards your SSH agent to remote machines", 9 November 2026, 63 points.. Canonical URL: https://www.realproblem.ai/idea/how-do-i-stop-my-ai-agent-from-using-my-ssh-keys-on-remote-machines.

How do I stop my AI agent from using my SSH keys on remote machines?

Desktop AI agents that work on remote machines can forward the developer's local SSH agent, so the agent can authenticate as them and push to private GitHub repos without any prompt or warning.

Who has it: Developers who connect ChatGPT Desktop, Codex or similar agent apps to remote servers, dev containers or cloud boxes over SSH.

Heard in public

“The agent running on a remote machine can authenticate using your local SSH credentials. That means it can push to your private repos on GitHub, for example.”

Quoted word for word from the public post linked below. Nobody submitted it to Real Problem AI.

Hacker News thread "ChatGPT Desktop forwards your SSH agent to remote machines", 9 November 2026, 63 points.

What nothing does yet

A credential guard for AI agents that shows which keys and tokens each agent session can reach, blocks SSH agent forwarding by default, and asks for a one-tap approval before an agent uses a key to sign in, push or deploy, with a log of every use.

Tools that solve this

3 named so far, and where each one stops.

  • SSH keys in the macOS Secure Enclave with a fingerprint per use

    Commenters suggest it so no new SSH session starts without a touch, but it takes manual setup and slows normal work.

  • Turning off agent forwarding in your SSH config

    Works if you know about it, but agent tools and dev container tools often write their own settings into your config.

  • Separate deploy keys per repo

    Limits what one key can reach, but each repo and machine needs its own key to be created and managed by hand.

How you might build it

  1. 01SSH agent proxy that approves each signing request
  2. 02Audit of SSH config changes made by agent and dev container tools
  3. 03Per-session scoped keys with short lifetimes
  4. 04Desktop approval prompt with a plain language reason
  5. 05Usage log of every key and token an agent touched

#NOV89 · Canonical URL: https://www.realproblem.ai/idea/how-do-i-stop-my-ai-agent-from-using-my-ssh-keys-on-remote-machines