Rendering an untrusted architecture-beta diagram lets the diagram author write an arbitrary property with the value horizontal or vertical onto Object.prototype. A group id of proto is accepted as a valid parent.

Any code in the same realm that reads a property of that name from an arbitrary object, or enumerates an object with bare for...in, observes the injected value (which can only be the string horizontal or vertical.

This may mean corrupted option/config defaults, bypassed truthiness checks, causing denial of service or logic corruption in the embedding application.

Because the injected value cannot be an object or function, this is not directly exploitable for remote code execution.

architecture-beta
   group mermaidPrototypePollutionMarker(cloud)[Marker]
   service a(server)[A] in  proto 
   service b(server)[B] in mermaidPrototypePollutionMarker
   a:R -- L:b

The vulnerable write was introduced in commit cb0a4703bdf01d47508bde1c08aa9a980d70bc20 and first shipped in mermaid@11.5.0. The lines are unchanged in every release since.

This has been patched by https://github.com/mermaid-js/mermaid/commit/99af3fc35ef0a9a9c8c6314521344d67523ddccf, released in Mermaid v11.16.1

There are no known workarounds. Please update to a patched version.

Are there any links users can visit to find out more?