Platform
javascript
Component
markdown-it-mermaid
Fixed in
0.15.3
CVE-2025-68669 is a critical Remote Code Execution (RCE) vulnerability affecting versions of 5ire AI Assistant up to and including 0.15.2. This vulnerability stems from an insecure configuration within the markdown-it-mermaid plugin, allowing attackers to inject malicious HTML. Successful exploitation could lead to complete system compromise. A patch is available in version 0.15.3.
The vulnerability lies in the useMarkdown.ts file where the markdown-it-mermaid plugin is initialized with securityLevel: 'loose'. This setting explicitly permits the rendering of HTML tags within Mermaid diagram nodes. An attacker can craft a malicious Mermaid diagram containing embedded HTML payloads. When this diagram is processed by 5ire, the injected HTML will be executed in the context of the application, granting the attacker arbitrary code execution capabilities. The potential impact is severe, ranging from data theft and modification to complete system takeover and potential lateral movement within the network. This vulnerability shares similarities with other HTML injection vulnerabilities where untrusted input is rendered without proper sanitization.
This vulnerability was publicly disclosed on 2025-12-23. No public proof-of-concept (PoC) code has been released at the time of publication, but the ease of exploitation suggests a high probability of exploitation. The CVSS score of 9.7 (CRITICAL) reflects the severity of the vulnerability. It is not currently listed on the CISA KEV catalog.
Exploit Status
EPSS
0.07% (22% percentile)
CISA SSVC
CVSS Vector
The primary mitigation is to upgrade to version 0.15.3 or later, which contains the fix. If upgrading is not immediately feasible, consider implementing input sanitization on all Mermaid diagrams before they are processed by 5ire. Specifically, restrict the allowed HTML tags and attributes within the diagrams. Another mitigation is to configure the markdown-it-mermaid plugin with a stricter securityLevel, such as 'strict' or 'safe', which disables HTML rendering. Additionally, review and restrict access to the 5ire application to limit the potential attack surface. After upgrading, confirm the fix by attempting to render a known malicious Mermaid diagram and verifying that the HTML is not executed.
Update the `markdown-it-mermaid` dependency to a version that fixes the vulnerability. If a patched version is not available, avoid using the `securityLevel: 'loose'` configuration and consider other safer alternatives for rendering Mermaid diagrams.
Vulnerability analysis and critical alerts directly to your inbox.
CVE-2025-68669 is a critical Remote Code Execution vulnerability in 5ire AI Assistant versions up to 0.15.2, allowing attackers to execute arbitrary code via malicious Mermaid diagrams due to an insecure plugin configuration.
If you are using 5ire AI Assistant version 0.15.2 or earlier, you are affected by this vulnerability. Upgrade to version 0.15.3 to mitigate the risk.
The recommended fix is to upgrade to version 0.15.3. As a temporary workaround, sanitize Mermaid diagram input and configure the markdown-it-mermaid plugin with a stricter securityLevel.
While no active exploitation has been confirmed, the vulnerability's severity and ease of exploitation suggest a high probability of exploitation.
Refer to the 5ire security advisories page for the latest information and official guidance regarding CVE-2025-68669.
Upload your dependency file and we'll tell you instantly if this and other CVEs hit you.