BestWebExtension Guide

How to Check a Chrome Extension's Permissions

A permission list is the only proof you get before installing. Read the install prompt, the store's privacy tab and the details page, and know what they still miss.

An extension install starts with a permission list, and almost everyone clicks past it. The list is short, and short is easy to misread in both directions. It can look harmless while the extension holds more than the prompt shows. It can look alarming when the access is the whole point of the tool.

Chrome gives you three places to look: the install prompt, the privacy tab on the store listing, and the extension’s details page after install. Reading them takes about two minutes.

What you need

1. Read the install prompt as a list of warnings

When you press Add to Chrome and the dialog names permissions, you are reading Chrome’s warnings, not the extension’s full manifest. Chrome draws a warning line for the permissions that can affect privacy, and shows them before installation or at runtime.

The gaps matter. storage, contextMenus and activeTab carry no warning at all, so an extension can hold all three without a single line in the prompt. Chrome’s advice to developers is to use optional permissions wherever a feature allows it, which keeps the audible list short.

So a quiet prompt tells you the extension avoids the loud APIs. It does not tell you how much of the web it can reach. That is the next thing to check.

2. Separate what it can do from which sites it can do it on

The manifest has two axes. permissions names APIs such as tabs or downloads. host_permissions and the matches in content_scripts name sites, written as match patterns.

Scope shows up in the pattern. One domain is narrow. https://*/* or *://*/* is every site you visit, and that is where the warning “Read and change all your data on all websites” comes from.

Host access is also the part that can arrive late. optional_host_permissions are granted by the user at runtime instead of at install time, so a calm prompt can be followed by a request for a specific site when you first use the feature. Later is not worse. It is just a different place to look.

For a page reader or a screenshot tool, host access is the job. The question is whether the scope matches it. Every site is a wide ask for a tool that only acts where you click.

3. The permission that keeps an extension narrow

activeTab is the one to look for in anything that acts on the page you are reading. It gives the extension temporary access to the current tab when you invoke it, and the grant ends when you navigate away or close the tab. Chrome’s docs call it an alternative to <all_urls> for many uses, and it installs with no warning.

Restricted pages, including the browser’s own chrome:// addresses, are never included.

A formatter, a cleaner or a picker that only ever touches the page in front of you can usually ship with activeTab. When a tool like that asks for all sites instead, that is the question worth asking before you click through.

4. Read the privacy tab on the store listing

Before publishing, a developer fills in a privacy section: a single purpose, a justification for each permission, a declaration about remote code, and a data usage disclosure with a limited use certification. Chrome shows those disclosures to users.

Read it next to the permission list, not on its own. A formatter that reads every page while declaring it collects browsing history is telling you something about itself. A tool that reaches all sites and declares it collects nothing is making a claim you cannot verify from the store alone.

The tab records what the developer declared. Chrome reviews listings against their code, and a declaration is still a declaration.

5. Open the details page and narrow the site access

After install, the details page is the full list. Right-click the extension’s icon and choose Manage Extension, or open chrome://extensions and press Details.

Two switches there start off and only you can turn them on: access to file:// URLs and access in incognito mode. Chrome’s docs describe both on that page. Right-click the icon, choose Manage Extension, and scroll down to enable access to file URLs or incognito mode.

Site access is the control worth using. Chrome’s help page places it next to “Allow this extension to read and change all your data on websites you visit”, with the choices “On select”, “On specific sites” or “On all sites”. Narrow it to a click where the tool still works that way.

The same page has a limit worth knowing. Changing site access only affects extensions whose access comes from host permissions. An extension that changes lower-level network traffic through a VPN or proxy setting is not affected at all.

6. The warnings worth slowing down for

These are Chrome’s warning strings, from the permissions reference, each with the permission it belongs to.

Chrome’s docs are plain about why any of this exists: permissions limit the damage if an extension is compromised.

7. Check the manifest when the source is public

An open source extension lets you read the declaration itself. The manifest lists permissions, optional_permissions, host_permissions, optional_host_permissions and the matches of every content script. That is the same set you have been reading in the store, in the file that decides what the code is allowed to call.

Manifest V3 extensions are not allowed to run remotely hosted code, and the store asks developers to declare whether they use remote code at all. A listing that answers no should have a repository you can line up against the manifest. The extensions from BestWebExtension are open source and lean on optional permissions, which is the shape to look for.

To build one to that standard, How to Develop a Chrome Extension covers the manifest from the other side.

How to check the result

Line up the three lists: the prompt, the permission list on the details page, and the data disclosure. They should tell one story. Permissions on the details page that never reached the prompt are normal, because only the warning-bearing ones are shown.

The lists are also where the two extension roundups on this site start. Best Chrome Extensions for Web Developers picks tools by what they need access to, and Best JSON Viewer Extensions for Large Payloads looks at one job done at very different scales.

When it won’t help

Takeaways

Questions

Can a Chrome extension read a page without asking?

Only with a grant. If it has the activeTab permission, access starts when you click the extension and ends when you leave the page. Without host permissions or activeTab it cannot read page content at all.

Does a short install prompt mean the extension asks for nothing?

No. Permissions such as storage, contextMenus and activeTab show no warning, so they never appear in the prompt. The details page after install lists everything the extension holds, host access included.

Can I take a permission back later?

Host access, yes. On the details page you can narrow site access to a click, to specific sites or to all sites, and you can turn file URL and incognito access off. Permissions declared in the manifest cannot be removed by the user.

BestWebExtension BestWebExtension docs