A browser extension runs inside the program where you do your banking, your email and most of your work. That makes the question "what can this extension see?" more important than it is for most software, and the answer is easier to find than people expect. Every extension has to declare what it wants in a plain text file.
This guide explains how to read that declaration, what the common permissions mean, and how to install an extension from a ZIP file in Chrome or Edge when it is not listed in a store.
Where the permissions are declared
Every Chrome and Edge extension contains a file called manifest.json. It is the extension's own statement of what it is and what it needs. If you have the extension as a ZIP, unpack it and open that file in any text editor. If the extension is in a store, the listing shows a summary of the same information before you install.
The parts worth reading are:
permissions: the browser features the extension can use, such as tabs, storage or downloads.host_permissions: the websites whose pages or data it can reach.content_scripts: code the extension injects into web pages, and which sites it is injected into.background: code that can run when you are not interacting with the extension.
You do not need to be a developer to read this. You are looking for whether the access requested matches the job the extension claims to do.
What common permissions actually allow
storagelets the extension save its own settings. It is low risk and almost every extension has it.tabslets the extension see the address and title of your open tabs, and close, move or create tabs. A tab manager needs it. A colour picker does not.tabGroupslets it create and change tab groups.activeTabgives temporary access to the current page only when you click the extension. It is a much narrower grant than access to all sites.host_permissionsfor a specific site limits the extension to that site, for example its own update server.host_permissionsfor all sites (written as<all_urls>or*://*/*) allows the extension to read and change every page you visit. Ad blockers and password managers need this. Many other extensions ask for it and do not.- Content scripts on all sites have the same reach: the extension's code runs inside every page.
The single most useful check is that last pair. An extension with no content scripts and no broad host permissions cannot read what is on the pages you visit.
A worked example
Our own Tab Cleaner extension closes duplicate tabs and groups tabs by website. Its product page lists each permission beside the reason for it:
tabs, to read open tab addresses so duplicates can be found and closed, used only after you click the toolbar button.tabGroups, to create and remove the groups it makes.storage, to keep your settings on your machine.- Two host permissions, both for our own domain: one for the update check and one for anonymous usage counts, which you can switch off in the popup.
It requests no access to other websites and has no content scripts, so it cannot read page content. That is the shape to look for: every permission maps to a feature you can see, and nothing is left over. If an extension's listing does not explain its permissions, you are entitled to ask why not.
Questions to ask before you install
- Does each permission match a feature? Anything unexplained is a reason to pause.
- Does it need all sites? If so, is that essential to what it does?
- Who publishes it? Look for a real website, a contact address and a privacy policy.
- What does it send anywhere? A privacy statement should say what leaves your browser, not only that privacy matters.
- Do you still use it? Review your installed extensions a few times a year and remove the ones you have stopped using. Extensions can change owners, and an update can ask for more than the version you first approved.
Installing from a ZIP in developer mode
Extensions are normally installed from the Chrome Web Store or Microsoft Edge Add-ons. Some are distributed as a ZIP instead, often because a store listing is not live yet. You can load one yourself:
- Download the ZIP and unpack it to a folder you will keep. The browser loads the extension from that folder every time it starts, so do not unpack it into Downloads and delete it later.
- Open the extensions page. Go to
chrome://extensionsin Chrome, oredge://extensionsin Edge. - Turn on Developer mode. The switch is at the top right in Chrome and at the bottom left in Edge.
- Choose Load unpacked and select the unpacked folder, the one that directly contains
manifest.json. - Pin the extension from the extensions menu so its button stays on the toolbar.
If the browser reports that the manifest is missing, you have selected the folder one level too high or too low.
What is different about an unpacked extension
Loading an extension this way skips the store, and that has consequences you should know about:
- There is no store review. You are relying on your own check of the publisher and the manifest, which is why the steps above matter more here.
- It does not update automatically. To update, download the new version, replace the folder's contents and select the reload button on the extensions page.
- The browser may remind you that developer mode extensions are active. That is expected.
- Verify the download if you can. If the publisher prints a SHA-256 checksum for the ZIP, compare it before you unpack.
- Managed computers may block it. On a work machine, developer mode is often disabled by policy. Ask IT instead of looking for a way around it.
The short version
Read the permissions, check that each one has a purpose you can name, be wary of anything that wants every website, and prune what you no longer use. Two minutes with manifest.json tells you more about an extension than its star rating does.