Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
CVE-2026-39987_POC — Proof-of-concept exploit for CVE-2026-39987, a pre-authentication RCE in Marimo's /terminal/ws WebSocket endpoint that yields an interactive shell without credentials. | Kitploit
Tools/GitHubGitHub/julichaan/cve-2026-39987_poc
Vulnerability AnalysisExploitationWeb Application ExploitationPenetration TestingAuthenticationRemote Access Tool
GitHubjulichaan/cve-2026-39987_poc

CVE-2026-39987_POC

Proof-of-concept exploit for CVE-2026-39987, a pre-authentication RCE in Marimo's /terminal/ws WebSocket endpoint that yields an interactive shell without credentials.

View Repository
422 months agoNot yet reviewed
Share

PoC — CVE-2026-39987 · Marimo Pre-Auth RCE

CVSS 4.0: 9.3 (Critical) · AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H Affects: marimo <= 0.20.4 · Patched: 0.23.0 · CWE-306 (Missing Authentication) Advisory: GHSA-2679-6mx9-h9xc · Fix: marimo-team/marimo@c24d480

Root cause

The WebSocket endpoint /terminal/ws (marimo/_server/api/endpoints/terminal.py) only checks the session mode (EDIT) and platform support before accepting the connection and calling pty.fork(). It carries neither a @requires(...) decorator nor a validate_auth() call.

Contrast measured in 0.20.4:

  • 73 uses of @requires(...) across 18 endpoints.
  • /ws (notebook) -> 2 calls to validate_auth().
  • terminal.py -> 0 @requires, 0 validate_auth. <-- the flaw.

The AuthBackend middleware protects HTTP and the /ws websocket, but /terminal/ws accepts the connection with no cookie and no token and forks a PTY shell with the privileges of the marimo process.

Exploitation requirements

  • marimo started in EDIT mode (marimo edit ...), which is the editor's default mode.
  • A platform with pty support (Linux/macOS; Windows and Pyodide are not affected).
  • Network access to the port. No credentials required.

Attack scenario

An attacker who can reach a marimo edit server over the network (common in remote / container deployments, often running as root) gets an interactive shell without authenticating: credential theft (cat .env), lateral movement, malware deployment. Exploited in the wild < 10 h after disclosure; added to CISA's KEV catalog on 2026-04-23.

PoC usage

pip install "marimo==0.20.4" websockets
marimo edit demo.py --headless --port 2718 --token --token-password SECRET
python exploit.py ws://127.0.0.1:2718 "id; cat /etc/passwd"

exploit.py sends no token, no cookie and no Authorization header. It sends the command as raw text + "\r" (Enter), exactly as a user would type it in the terminal.

Result obtained in the lab (Kali, marimo 0.20.4)

  • Baseline WITHOUT token: HTTP -> 401 ; /ws -> 403 (auth DOES work).
  • /terminal/ws WITHOUT token -> ACCEPTED. Executed on the server: uid=1000() gid=1000() ... Linux ... and wrote /tmp/cve_2026_39987_proof.txt -> RCE confirmed.

Remediation

Upgrade to marimo >= 0.23.0. Temporary mitigation: do not expose marimo edit to untrusted networks; put the editor behind an authenticated proxy; run it unprivileged / in a container without secrets.

Disclaimer

For authorized security testing and educational use only. Run it only against systems you own or have explicit permission to test.

Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    916
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    558
  • 3

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    370
  • 4

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    360
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    317
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    289

Explore all tools

Browse our collection of tools

View all tools →