Skip to content

Security: OpenDeernet/opsnest

Security

SECURITY.md

Security notes

OpsNest is an Alpha desktop application. This document records the current release posture so users can make an informed decision before connecting real servers.

Current protections

  • Server credentials and AI API keys are stored through the operating-system credential store.
  • The local data file does not intentionally contain plaintext server passwords, private keys, or AI API keys.
  • Server command output is kept in local logs, with basic redaction before conversation logging.
  • AI and mutating operations pass through local application logic; the model does not receive SSH credentials.
  • The app does not require an OpsNest cloud backend.
  • SSH server keys use trust on first use (TOFU) with the user's standard ~/.ssh/known_hosts file. A changed key is rejected on later connections.

Known Alpha limitations

  • The first SSH connection automatically records the presented host key. Verify a new server's fingerprint through another trusted channel before entering credentials. The app does not yet provide a fingerprint confirmation or known-hosts management screen.
  • The current SSH stack supports RSA private keys through the RustCrypto rsa crate. RustSec advisory RUSTSEC-2023-0071 has no patched release; prefer password or Ed25519-key authentication until the upstream dependency provides a fix.
  • The Agent can still generate Shell commands. Review the plan and output before approving changes.
  • Server output and selected log/configuration content may be sent to the model provider configured by the user.
  • The app has been exercised mainly on Windows x64 and a limited set of Linux-based systems.

Before production use

Use a test server or a verified backup. Do not use an unreviewed Alpha build for unattended production changes. Before the stable release, the project should add a user-facing host-fingerprint workflow, stronger redaction tests, a documented threat model, and a dependency vulnerability scan in CI.

Reporting a vulnerability

Please do not publish credentials or an exploitable proof of concept in a public Issue. Report a suspected vulnerability privately through the repository's private contact channel, or open a minimal Issue asking for a private contact method. Include the affected version, operating system, reproduction steps, and impact, with all addresses, tokens, passwords, and logs redacted.

There aren't any published security advisories