tc4dy/CVE-2026-6875-PoC-Exploit

GitHubRepository PoCExploit

PoC files

7 files

File viewing is interactive and short-lived. Downloads are password-protected ZIP archives using password eip.

GitHub
Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a full weaponized exploit framework for CVE-2026-6875, a pre-auth RCE in ServiceNow AI Platform. It chains JavaScript injection with a sandbox escape to execute arbitrary system commands, and includes post-exploitation modules for privilege escalation, persistence, lateral movement, and credential dumping.

Backdoor review

No backdoor observed in reviewed code

The repository contains a disclosed exploit framework and verifier for CVE-2026-6875, a ServiceNow pre-auth RCE. The Python code in exploit.py implements the described JavaScript injection, sandbox escape, and post-exploitation modules (shell, privesc, persistence, lateral movement, etc.) against a target ServiceNow instance. All behavior is consistent with the advertised exploit functionality and operates against the operator-supplied target. No concealed payloads, unauthorized data exfiltration, or deceptive local execution against the operator were found.

ClassificationExploit
Model confidence98%
AuthenticationNot required
LanguagesPython
Target softwareServiceNow AI Platform
Attack typesRemote Code ExecutionPrivilege EscalationPersistenceLateral MovementCredential Dumping
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact's primary purpose is to exploit CVE-2026-6875 for remote code execution. The code constructs and delivers a JavaScript injection payload that escapes the sandbox and executes arbitrary system commands via java.lang.Runtime.exec(). It includes an interactive shell, reverse shell, webshell upload, and multiple post-exploitation modules. Although a --detect flag exists, the overwhelming majority of the codebase is dedicated to exploitation, not mere detection.

exploit.py:92-107exploit.py:109-133exploit.py:272-313exploit.py:451-515README.md:1-2README.md:70-72

Requirements

  • Target must be a vulnerable, self-hosted ServiceNow AI Platform instance with the assessment_thanks.do endpoint exposed.README.md:48README.md:63exploit.py:190
  • Attacker must have network access to the target's HTTPS port (default 443).exploit.py:190exploit.py:653

Observed behavior

  • Sends a POST request to /assessment_thanks.do with a crafted sysparm_assessable_type parameter containing a JavaScript injection payload.exploit.py:190-208exploit.py:92-107
  • The injected JavaScript escapes the ServiceNow script sandbox using a gadget chain involving DiscoveryFunctions, AbstractAjaxProcessor, and Class.create, then executes arbitrary Java code via java.lang.Runtime.exec().exploit.py:92-107exploit.py:109-133README.md:256-273
  • Provides an interactive shell for executing arbitrary OS commands, uploading webshells, downloading files, and triggering post-exploitation modules (privesc, persistence, lateral movement, credential dumping, clean tracks).exploit.py:451-515exploit.py:315-330exploit.py:411-440
  • Includes a detection mode (--detect) that checks for vulnerability by sending a benign echo command and inspecting the response, but the primary purpose is exploitation.exploit.py:247-270README.md:84
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Url
https://github.com/tc4dy/CVE-2026-15409-15410-FrameworkLink to another exploit repository by the same author, listed in README.md references.README.md:346
Url
https://github.com/tc4dy/CVE-2026-41091-PoC-ExploitLink to another exploit repository by the same author, listed in README.md references.README.md:349
Url
https://github.com/tc4dy/CVE-2026-57821-PoC-ExploitLink to another exploit repository by the same author, listed in README.md references.README.md:352
Url
https://github.com/tc4dy/CVE-2026-41940-PoC-ExploitLink to another exploit repository by the same author, listed in README.md references.README.md:355
Review boundaries

What the analysis did not establish

  • The evidence includes only three text files (README.md, exploit.py, requirements.txt) out of seven total files; four files (including one non-text media file) are omitted, so the analysis may not cover all repository contents.
  • The verifier.py file mentioned in the README is not included in the supplied evidence, so its behavior cannot be confirmed.
  • The analysis is based solely on static code review; the code was not executed, and its actual effectiveness or safety is not verified.
  • The repository contains 4 files not included in the text evidence (1 non-text media file, 3 unclassified files), which were not analyzed. Their content could theoretically contain concealed behavior, but no positive evidence suggests this.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

Linked vulnerabilities

1