Critical Marimo Pre-Auth RCE Actively Exploited

A newly disclosed vulnerability in the Marimo open-source reactive Python notebook platform is now under active exploitation, just hours after its public disclosure. Tracked as CVE-2026-39987, the flaw allows unauthenticated remote code execution via the /terminal/ws WebSocket endpoint, exposing an interactive shell without proper authentication checks.

Vulnerability Details

• Affected versions: Marimo 0.20.4 and earlier.
• Severity: Critical (CVSS 9.3).
• Root cause: /terminal/ws endpoint exposes an interactive terminal without authentication.
• Impact: Attackers gain direct shell access with the same privileges as the Marimo process.

Exploitation in the Wild

• Timeline: Exploits observed within 10 hours of disclosure.
• Reconnaissance: 125 IPs began scanning within 12 hours.
• Attack behavior:
  • Initial validation with simple commands (pwd, whoami, ls).
  • Credential harvesting from .env files, environment variables, and SSH keys.
  • Focused on cloud credentials and application secrets.
• Operator profile: Manual, methodical attacker — not automated scripts.
• Goal: Quick credential theft rather than persistence or cryptomining.

Mitigation Guidance

• Upgrade immediately: Patch to Marimo 0.23.0.
• Restrict access: Block external connections to /terminal/ws via firewall.
• Rotate secrets: Replace any exposed credentials, especially cloud keys and SSH.
• Monitor activity: Watch for suspicious WebSocket connections and credential access attempts.
• Fallback mitigation: If upgrading isn’t possible, disable or block /terminal/ws entirely.

Final Thought

This incident underscores the speed at which attackers weaponize newly disclosed vulnerabilities. For open-source projects like Marimo, disclosure-to-exploitation windows are shrinking to mere hours. Organizations must adopt rapid patching, secret rotation, and proactive monitoring to defend against opportunistic credential theft.