Active extension probe (AED)
Fetches a hidden file from each of thousands of known extension IDs to learn which ones are installed.
In one line
Can check which of thousands of known browser extensions you have installed
Fetches a hidden file from each of thousands of known extension IDs to learn which ones are installed.
Reveals installed add-ons, which can imply job search, sales role, religion, health or politics, and a stable identifier for the browser profile. LinkedIn calls it abuse prevention; critics call it covert profiling.
What it can and can't do
What the user does: Nothing. It happens when you open the site in Chrome or Edge.
It can
- Test a long list of known extension IDs, one by one
- Tell which of those extensions are installed
- Use the result as a stable fingerprint of your browser profile
- Send the list of detected extensions to its servers
It cannot
- See extensions that are not on its list
- See what an extension does or the data inside it
- Work in Firefox, Safari or most phone browsers
- Find extensions that keep their files hidden from web pages
How it differs between platforms: Works in Chrome and Edge only. Firefox, Safari and phone browsers block it.
Browser support
| Browser | Status | First version | Released | Note |
|---|---|---|---|---|
| Chrome | Supported | Not available | Not listed | Needs chrome-extension:// scheme |
| Edge | Supported | 79 | 2020-01-15 | Chromium Edge: UA contains Chrome, extension IDs from the Chrome Web Store work if installed there |
| Firefox | Not supported | Not available | Not listed | moz-extension URLs use per-install random UUIDs, so IDs cannot be guessed |
| Safari | Not supported | Not available | Not listed | Safari extension URLs are per-install random |
| Safari on iPhone | Not supported | Not available | Not listed | None |
| Chrome on Android | Not supported | Not available | Not listed | Chrome for Android has no extension support |
| Samsung Internet | Not supported | Not available | Not listed | None |
How a site asks for it
| API | fetch('chrome-extension://<id>/<file>') per list entry, Promise.allSettled; fulfilled = installed |
| How access is granted | none |
| What the user sees | None (no browser prompt) |
| Needs a secure context (HTTPS) | No |
| Needs a user gesture | No |
| How long it lasts | Runs on page load for every visit; result is sent once per run if any ID is found |
| How to take it back | Not listed |
| Inside an iframe | Not listed |
Typical flow
- Script checks navigator.userAgent contains 'Chrome' and that it is not server side
- For each hardcoded {id, file} pair, fetch chrome-extension://id/file
- A fulfilled fetch means the extension is installed and exposes that file in web_accessible_resources
- Collected IDs are sent as AedEvent.browserExtensionIds through the tracking payload, only if the list is non-empty
- Later versions add staggered scanning, requestIdleCallback and a 2000 ms timeout (per BrowserGate)
Data you get
What comes back: Set of installed extension IDs out of the probe list
Shape: list
How often: every page load or session start
| Field | Type | Unit | Range | Precision |
|---|---|---|---|---|
browserExtensionIds | list | extension IDs | 0 to list size (2,953 in the public snapshot, 6,236 claimed Apr 2026) | 32 char lowercase ID each |
probe_list_size | integer | IDs | 2,000 (2025 gist) to 6,236 (Apr 2026 claim) | exact |
Technical limits
- Only Chromium browsers with chrome-extension:// URLs
- Only finds extensions that list a file in web_accessible_resources
- Manifest V3 use_dynamic_url randomizes the URL so those extensions cannot be probed
- Cannot read extension data, only presence
The story
What it is
A site quietly tests thousands of known browser add-on addresses to learn which add-ons you have installed.
For you
You see nothing at all. While the page loads, it tries thousands of hidden add-on addresses in the background. The answer is a list of which add-ons your browser has.
For makers
The site says this catches scraping and automation tools that break its rules. Critics say the same list can hint at your job, beliefs or health, and acts as a steady ID for your browser.
Where you see it
| Type | Example | Apps |
|---|---|---|
| security | A professional network checks for known scraper add-ons when a profile gets viewed hundreds of times in a minute, then limits the account. | |
| security | A sneaker or concert ticket site looks for auto-buyer add-ons before a big drop and sends those visitors to extra checks. | None named |
| games | A browser game with real prizes looks for auto-clicker or cheat add-ons before a ranked match starts. | None named |
| ads | An ad network adds the add-on list to its other clues so it can tell a returning browser from a new one. | None named |
| analytics | A fraud team compares add-on lists across thousands of new accounts and finds a cluster of fake profiles that all share the same odd set. | None named |
| enterprise | A sales software maker spots a rival's sales add-on in a visitor's browser, which critics call competitor snooping. | |
| social | A platform could guess that someone is job hunting from a resume or job alert add-on, which privacy groups warn about. | None named |
Risks
- Reveals private traits
- Hidden stable ID
- No prompt or consent
Good uses and abuse
Legitimate uses
- Detect scraper and automation extensions that abuse the site
- Block known data harvesting tools
Known abuse
- Profiling users by installed extensions (religion, health, job hunting, rivals' tools)
- Competitor tool detection
- Stable fingerprint component
Scores
| Score | Value | Meaning |
|---|---|---|
| Sensitivity | 4 of 5 | How revealing the data is |
| Utility | 3 of 5 | How useful it is to honest apps |
| Friction | 0 of 5 | How much the user has to do |
| Fingerprint bits | 12 | Rough identifying entropy, when known |
A real example: LinkedIn
Core of BrowserGate. LinkedIn script loops a hardcoded extension ID list (2,953 IDs in the public mdp snapshot, 6,236 claimed Apr 2026) and fetches each chrome-extension:// resource. Verified in the snapshot; LinkedIn says it is for abuse prevention.
History
- 2025, Chrome: About 2,000 IDs listed in a public gist (cited by BleepingComputer)
- 2026-04, Chrome: BrowserGate reports 6,236 IDs; public snapshot has 2,953
Related capabilities
Related
- Spectroscopy DOM scan for extension traces. Can look through its own page for traces that extensions leave behind
- APFC / DNA device fingerprint (abuse prevention features). Can collect about 48 device traits and send them in secret with later requests
- User agent string. Can see your browser, version, and operating system from the text it sends
Sources
- mdp/linkedin-extension-fingerprinting (LinkedIn script snapshot and ID list)
- BrowserGate: how it works (48 feature table)
- BleepingComputer: LinkedIn scans for 6,000+ Chrome extensions
- Chrome docs: web_accessible_resources
- Ars Technica via mirror: LinkedIn scanning sparks controversy
How well checked: Checked against real behavior.