CVE-2026-25253
OpenClaw WebSocket Token Disclosure Vulnerability
Description
OpenClaw (aka clawdbot or Moltbot) before 2026.1.29 obtains a gatewayUrl value from a query string and automatically makes a WebSocket connection without prompting, sending a token value.
INFO
Published Date :
Feb. 1, 2026, 11:15 p.m.
Last Modified :
Feb. 13, 2026, 5:41 p.m.
Remotely Exploit :
Yes !
Source :
[email protected]
CVSS Scores
| Score | Version | Severity | Vector | Exploitability Score | Impact Score | Source |
|---|---|---|---|---|---|---|
| CVSS 3.1 | HIGH | 8254265b-2729-46b6-b9e3-3dfca2d5bfca | ||||
| CVSS 3.1 | HIGH | [email protected] |
Solution
- Update OpenClaw to version 2026.1.29 or newer.
- Review and secure gatewayUrl configurations.
- Implement user consent for WebSocket connections.
Public PoC/Exploit Available at Github
CVE-2026-25253 has a 83 public
PoC/Exploit available at Github.
Go to the Public Exploits tab to see the list.
References to Advisories, Solutions, and Tools
Here, you will find a curated list of external links that provide in-depth
information, practical solutions, and valuable tools related to
CVE-2026-25253.
| URL | Resource |
|---|---|
| https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys | Exploit Third Party Advisory |
| https://ethiack.com/news/blog/one-click-rce-moltbot | Exploit Third Party Advisory |
| https://github.com/openclaw/openclaw/security/advisories/GHSA-g8p2-7wf7-98mq | Vendor Advisory |
| https://openclaw.ai/blog | Product |
| https://x.com/0xacb/status/2016913750557651228 | Exploit |
| https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys | Exploit Third Party Advisory |
CWE - Common Weakness Enumeration
While CVE identifies
specific instances of vulnerabilities, CWE categorizes the common flaws or
weaknesses that can lead to vulnerabilities. CVE-2026-25253 is
associated with the following CWEs:
Common Attack Pattern Enumeration and Classification (CAPEC)
Common Attack Pattern Enumeration and Classification
(CAPEC)
stores attack patterns, which are descriptions of the common attributes and
approaches employed by adversaries to exploit the CVE-2026-25253
weaknesses.
We scan GitHub repositories to detect new proof-of-concept exploits. Following list is a collection of public exploits and proof-of-concepts, which have been published on GitHub (sorted by the most recently updated).
The Brain Problem: A Security Report on OpenClaw's Skill Ecosystem
clawdbot moltbot openclaw etc... vuln scanner and exploiter
Python
> OpenClaw security audit and hardened deployment guide — known vulnerabilities (CVE-2026-25253, malicious skills, credential leakage), architectural mitigations, and a step-by-step VPS deployment plan
ai-agent ai-agents clawdbot cybersecurity deployment-guide llm-security openclaw security-audit self-hosted vps-deployment hardening-llm-security
None
None
Python
Security-hardened AI agent platform addressing OpenClaw/Moltbot vulnerabilities (CVE-2026-25253)
Dockerfile TypeScript JavaScript
🚫🦞 Fastest, tiniest, almost invisible AI assistant infrastructure with security-first mindset that can run anywhere, even in your browser...
agents ai openclaw zeroclaw allvibenocode noclaw
🛡️ Secure local AI assistant with gVisor/Kata Containers sandboxing and Human-in-the-Loop approval system
Dockerfile Python JavaScript Shell CSS TypeScript PowerShell
Multi-agent security platform for AI agent protection.
agents elasticsearch openclaw python docker
Python Shell
Security-first AI agent for WhatsApp. Built as an answer to OpenClaw's architecture.
ai-agent llm openclaw prompt-injection python security typescript whatsapp
Dockerfile Python TypeScript Shell HTML
Secure autonomous AI agent fleet platform — Docker-isolated, multi-provider, with built-in cost controls. OpenClaw alternative for production use.
agent-framework ai-agents ai-orchestration autonomous-agents llm-agents multi-agent openclaw-alternative self-hosted ai-runtime ai-runtime-security nanoclaw openclaw nanobot zeroclaw memu ai-assistant docker-sandbox
Python Makefile Shell PowerShell
🛡️ Secure local AI assistant with gVisor/Kata Containers sandboxing and Human-in-the-Loop approval system
Python JavaScript CSS TypeScript PowerShell Shell Dockerfile
None
Shell HCL
"A local-first, Rust-hardened security layer for AI agents. Neutralizing RCE and data theft in the OpenClaw ecosystem through hardware-bound attestation and encrypted memory shards.#ai-security#rust#openclaw#cybersecurity#sovereign-ai#zero-trust
Rust Shell
Tracking OpenClaw CVEs
Shell Python Jinja
Results are limited to the first 15 repositories due to potential performance issues.
The following list is the news that have been mention
CVE-2026-25253 vulnerability anywhere in the article.
-
CybersecurityNews
Multiple Hacking Groups Exploit OpenClaw Instances to Steal API key and Deploy Malware
Hacking Groups Exploit OpenClaw Instances A widespread exploitation of OpenClaw, formerly known as MoltBot and ClawdBot, by multiple hacking groups to deploy malicious payloads. OpenClaw, an open-sour ... Read more
-
CybersecurityNews
OpenClaw AI Framework v2026.2.17 Released with Anthropic Model Support and Security Fixes
OpenClaw AI Framework v2026.2.17 Released OpenClaw has released version 2026.2.17 with significant enhancements, including support for Anthropic’s Claude Sonnet 4.6 model. Expanded context windows, th ... Read more
-
Kaspersky
Key OpenClaw risks, Clawdbot, Moltbot | Kaspersky official blog
Everyone has likely heard of OpenClaw, previously known as “Clawdbot” or “Moltbot”, the open-source AI assistant that can be deployed on a machine locally. It plugs into popular chat platforms like Wh ... Read more
-
reddit.com
Architectural Isolation Tradeoffs in the OpenClaw Ecosystem After CVE-2026-25253
Let us know your cookie preferences Reddit uses cookies and similar technologies to: Keep the website operational and running properly Prevent fraud and abuse Monitor site usage and performance metric ... Read more
-
CybersecurityNews
15,200 OpenClaw Control Panels with Full System Access Exposed to the Internet
OpenClaw Control Panels Exposed A critical security failure in the rapidly adopting “agentic AI” ecosystem has left tens of thousands of personal and corporate AI assistants fully exposed to the publi ... Read more
-
The Register
More than 135,000 OpenClaw instances exposed to internet in latest vibe-coded disaster
It's a day with a name ending in Y, so you know what that means: Another OpenClaw cybersecurity disaster. This time around, SecurityScorecard's STRIKE threat intelligence team is sounding the alarm ov ... Read more
-
SentinelOne
The Good, the Bad and the Ugly in Cybersecurity – Week 6
The Good | Former Google Engineer Steals AI Supercomputing Secrets for China Former Google software engineer Linwei Ding has been found guilty of economic espionage and trade secret theft after steali ... Read more
-
SentinelOne
The Good, the Bad and the Ugly in Cybersecurity – Week 6
The Good | Former Google Engineer Steals AI Supercomputing Secrets for China Former Google software engineer Linwei Ding has been found guilty of economic espionage and trade secret theft after steali ... Read more
-
The Hacker News
OpenClaw Bug Enables One-Click Remote Code Execution via Malicious Link
A high-severity security flaw has been disclosed in OpenClaw (formerly referred to as Clawdbot and Moltbot) that could allow remote code execution (RCE) through a crafted malicious link. The issue, wh ... Read more
The following table lists the changes that have been made to the
CVE-2026-25253 vulnerability over time.
Vulnerability history details can be useful for understanding the evolution of a vulnerability, and for identifying the most recent changes that may impact the vulnerability's severity, exploitability, or other characteristics.
-
Initial Analysis by [email protected]
Feb. 13, 2026
Action Type Old Value New Value Added CPE Configuration OR *cpe:2.3:a:openclaw:openclaw:*:*:*:*:*:node.js:*:* versions up to (excluding) 2026.1.29 Added Reference Type MITRE: https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys Types: Exploit, Third Party Advisory Added Reference Type MITRE: https://ethiack.com/news/blog/one-click-rce-moltbot Types: Exploit, Third Party Advisory Added Reference Type MITRE: https://github.com/openclaw/openclaw/security/advisories/GHSA-g8p2-7wf7-98mq Types: Vendor Advisory Added Reference Type MITRE: https://openclaw.ai/blog Types: Product Added Reference Type MITRE: https://x.com/0xacb/status/2016913750557651228 Types: Exploit Added Reference Type CISA-ADP: https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys Types: Exploit, Third Party Advisory -
CVE Modified by 134c704f-9b21-4f2e-91b3-4a467353bcc0
Feb. 03, 2026
Action Type Old Value New Value Added Reference https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys -
CVE Modified by [email protected]
Feb. 02, 2026
Action Type Old Value New Value Added Reference https://ethiack.com/news/blog/one-click-rce-moltbot Added Reference https://x.com/0xacb/status/2016913750557651228 -
New CVE Received by [email protected]
Feb. 01, 2026
Action Type Old Value New Value Added Description OpenClaw (aka clawdbot or Moltbot) before 2026.1.29 obtains a gatewayUrl value from a query string and automatically makes a WebSocket connection without prompting, sending a token value. Added CVSS V3.1 AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H Added CWE CWE-669 Added Reference https://depthfirst.com/post/1-click-rce-to-steal-your-moltbot-data-and-keys Added Reference https://github.com/openclaw/openclaw/security/advisories/GHSA-g8p2-7wf7-98mq Added Reference https://openclaw.ai/blog