S4E just found a high-severity finding from top 10 tcp port service scan
PlatformPlansPartners
Resources
ToolsBlogDocs
Sign Up →
critical·Product Based Web Vulnerabilities·Updated Apr 30, 2026

CVE-2026-33057 Scanner

CVE-2026-33057 Scanner - Remote Code Execution (RCE) vulnerability in Mesop AI Sandbox

Est. Time~10 seconds
Scan TypeSingle Scan
Targetsdomain, subdomain, ipv4
CostFree
3.2k
Times Used
continuous scan runs
5.8k
Continuously Checked
assets under CS
0
Vulnerabilities Found
confirmed findings
References
CVECVE-2026-33057
9.8
CVSScritical
Exploitable remotely over the internet · no authentication required.

Mesop is a Python-based UI framework that allows users to build web applications. In versions 1.2.2 and below, an explicit web endpoint inside the ai/ testing module infrastructure directly ingests untrusted Python code strings unconditionally without authentication measures, yielding standard Unrestricted Remote Code Execution. Any individual capable of routing HTTP logic to this server block will gain explicit host-machine command rights. The AI codebase package includes a lightweight debugging Flask server inside ai/sandbox/wsgi_app.py. The /exec-py route accepts base_64 encoded raw string payloads inside the code parameter natively evaluated by a basic POST web request. It saves it rapidly to the operating system logic path and injects it recursively using execute_module(module_path...). This issue has been fixed in version 1.2.3.

Attack Vector
Network
Privileges Req.
None
User Interaction
None
Affected
mesopby mesop-dev
< 1.2.3
Updated Sep 9, 2026View on NVD →
Detail

Mesop AI Sandbox is utilized by developers and data scientists for running isolated testing environments, allowing them to execute AI models securely and with flexibility. It aims to provide a versatile platform for testing code and simulations in controlled conditions without risking the main system. By offering a sandbox environment, it enables developers to refine algorithms and adjust parameters in real-time. Companies that operate with AI models benefit from using Mesop AI Sandbox to ensure the robustness of their solutions before deployment. It is extensively adopted in research institutions where AI algorithms undergo frequent modifications and testing. The tool is primarily valued for its ease of use and the security it assures during developmental stages.

The vulnerability in question enables Remote Code Execution, which is considered a severe security risk due to its potential impact on system integrity. This occurs when constraints on code execution are bypassed, allowing attackers to inject and execute malicious code remotely. Such vulnerabilities typically arise from inadequate validation of user input or exposure of dynamic execution endpoints. Attackers can leverage this flaw to run arbitrary code on the affected system, which could lead to unauthorized access and control. The exploitation of this vulnerability does not require user interaction and can often be executed remotely over the network. This RCE vulnerability is noteworthy because it can compromise the underlying infrastructure by leveraging the unrestricted nature of the sandbox environment.

The technical aspect of this vulnerability lies within the Mesop AI Sandbox's /exec-py endpoint, used for testing Python code execution. Attackers can target this endpoint to pass base64-encoded malicious Python scripts, which the system ingests and executes without proper authentication checks. The "code" parameter in the HTTP POST request facilitates this exploitation, as it does not enforce proper validation or restrictions on the content being submitted. The use of base64 encoding facilitates the bypassing of simple input filters, thus increasing the likelihood of a successful attack. Triggered scripts can manipulate system files, extract sensitive data, and even disable system defenses. This specific security flaw exploits the design assumption that input reaching the endpoint would be from a trusted and authenticated source.

Exploiting the Remote Code Execution vulnerability may lead to severe consequences, such as complete system takeovers, data theft, and service disruptions. Attackers can gain root or administrative privileges, allowing them to persistently install backdoors or other malware. Once compromised, an affected system could be used as a launchpad for attacks on other network resources. Organizations facing such vulnerabilities risk losing sensitive internal data and customer information, damaging their reputation and trust. Additionally, the cost and time required to mitigate and recover from such breaches can be substantial. The effects also extend to regulatory non-compliance, especially if data subjects are involved, leading to potential legal ramifications.

REFERENCES

Solution Advice
  • Upgrade Mesop AI Sandbox to version 1.2.3 or a later version to close the vulnerability.
  • Ensure all endpoints require proper authentication and access controls to prevent unauthorized use.
  • Implement thorough input validation checks for all data being processed by the application, especially at execution points.
  • Regularly audit and test sandbox environments for potential security flaws that might arise from newly implemented code or changes.
  • Configure network and endpoint security solutions to monitor and block suspicious activities targeting known vulnerable endpoints.

Get AI-powered remediation steps tailored to your asset.

Try AI Solutions →

Check your infrastructure.
Right now.

11,000+ scanners. Free to start. No credit card required.