Trust, then read the manifest

Are Chrome Extensions Safe for Business? The Published Incidents, the Sold-Extension Problem and How a Build Limits Its Own Risk

On Christmas Eve 2024, according to Cyberhaven, a phishing attack compromised an employee's access to the Chrome Web Store. Nobody broke the extension's code. They borrowed the right to publish it, and the browser's automatic updates did the rest.

November 17, 2026 · 9 min read

acmedentalgroup.com · staff extension review
Extensions on the front desk browsers, read one by one
Two extensions ask for far more than their jobs need
  • PassInstalls limited to an approved list
  • Warn2-Step Verification with a security keyphone codes
  • FailPDF tool asks to read all websitesall sites
  • FailTab saver changed publisher last monthnew owner
  • PassScheduling helper uses activeTab only
  • WarnPayroll site blocked for extensions by policynot set
Illustrative: a sample extension review for a fictional dental group

The short answer

Chrome extensions can be safe for business use, but each one is software that updates itself and can reach whatever its permissions allow. Google's Chrome Security Team wrote in June 2024 that less than 1% of Chrome Web Store installs that year were found to include malware. The larger business risks are a phished developer account, an extension sold to a new owner, and permissions far broader than the job.

On this page
  1. Christmas 2024: one phished login, one poisoned update
  2. The extension that quietly changed hands
  3. What Google screens for, in its own words
  4. The phishing email dressed as the Chrome Web Store
  5. How a careful build shrinks its own blast radius
  6. For the IT side: deciding which extensions staff can run
  7. What we promise, and what nobody should

Christmas 2024: one phished login, one poisoned update

Cyberhaven described its own incident in a post dated December 27, 2024 and last updated January 2, 2025. The post's live address now redirects elsewhere, so we read it through an Internet Archive copy from January 15, 2025. Everything in this section is according to that post.

  • On December 24, "a phishing attack compromised a Cyberhaven employee's access to the Google Chrome Web Store," and the attacker used it to publish a malicious version, 24.10.4.
  • The malicious code was active from 1:32 AM UTC on December 25 to 2:50 AM UTC on December 26.
  • Cyberhaven says its team detected the compromise at 11:54 PM UTC on December 25 and removed the package within 60 minutes.
  • "Only Chrome-based browsers that auto-updated during this period were impacted," and the code "could have exfiltrated cookies and authenticated sessions for certain targeted websites."
  • A secure version, 24.10.5, was published and deployed automatically.
  • The post adds that "Public reports suggest this attack was part of a wider campaign to target Chrome extension developers."

The lesson is not that Cyberhaven wrote bad code. The code was never the weak point. The publishing account was, and the same automatic update that delivers a fix in hours delivered the attack in hours too. Any business that ships an extension, or relies on one, owns that risk.

The extension that quietly changed hands

The second risk needs no attacker at all. On June 19, 2020, the original developer of The Great Suspender extension posted on the project's GitHub page that a new owner had "purchased the rights to publish the extension to the Chrome webstore" and would manage the public release process going forward. After that, updates to everyone who had installed it would come from the new owner.

By early February 2021 the extension's Chrome Web Store listing was gone, according to an Internet Archive capture from February 6, 2021 that shows a missing page. We found no Google statement giving a reason, so we give none. The point stands without one: an extension can be sold, the buyer gets the update channel, and the people using it are not asked.

What Google screens for, in its own words

Google's Chrome Security Team described its checks in a post on Google's security blog, Staying Safe with Chrome Extensions, dated June 20, 2024. It names an automated review by "machine-learning systems" and "A human review," and says Google's systems keep watching after approval, "periodically reviewing what extensions are actually doing and comparing that to the stated objectives." The same post, still published as of November 2026, gives Google's figure: "In 2024, less than 1% of all installs from the Chrome Web Store were found to include malware."

Google's protections as its documentation describes them, as of November 2026, and what each leaves to you
What Google says it doesWhat it does not cover
Review before publishingManual and automated review of every submission, new or updatedA malicious update from a stolen account still has to be caught in review
Ongoing checksExisting items are reviewed periodically for complianceThe window between an update and the next check
Malware takedown"The violating extension is disabled on all end user devices" and cannot be re-enabledData taken before the takedown
Remote code banManifest V3 extensions run only JavaScript in their reviewed packageWhat the reviewed code itself sends to a server
Account security2-Step Verification required before publishing or updatingA developer who approves a phishing request anyway

Sources: Chrome's review process page, the Manifest V3 overview and the two-step verification policy. One detail on the review page surprises developers: when a malware takedown happens, "notification is not sent to the publisher's email address." The first sign may be customers asking where the extension went.

Dark green chalkboard with chalk arrows linking doodled browser windows to a padlock, a black-and-white halftone cutout of an old key in front
Google's checks are one layer; the developer account and the permission list are the others.

The phishing email dressed as the Chrome Web Store

Cyberhaven was not the only developer targeted. In a thread titled "Beware Phishing in Chrome Web Store." in the Chromium Extensions group, a Chrome Web Store team post on January 6, 2025 said that "developers were tricked into granting oauth permission to the chrome web store api" and that affected apps "are disabled as soon as they are discovered." According to that post, "the Chrome Web Store team will not be asking someone to grant us access to their account by our own API." Any email asking for exactly that is the attack.

Chrome's security guidance for developers states the stakes in one sentence: compromised accounts let an attacker "push malicious code directly to all users." Its advice is two-factor authentication, "preferably with a security key." Here is the account hygiene we set up for clients:

  1. 1

    Own the account as a company

    Open it on an address the business controls. Google's registration page says the email cannot be changed after the account is created.

  2. 2

    Use a security key for the second factor

    Google requires 2-Step Verification; a physical key is what its security page prefers.

  3. 3

    Grant nobody API access by email

    Approve OAuth requests only from flows you started yourself, never from a link in a message.

  4. 4

    Lock away signing files

    If you pack the extension locally, Google says to keep the .pem file "in a secret and secure place" because updates need it.

  5. 5

    Know who can publish

    Keep the list of people with publishing rights short, and remove access the day someone leaves.

How a careful build shrinks its own blast radius

No developer can make an extension that cannot be hacked, and we will not claim to. What a build can do is limit what a hijacked copy could reach. Chrome's permissions guide says it directly: "Permissions help to limit damage if your extension is compromised by malware."

Short permission list

Every permission is something a hijacked update inherits. Google's policy asks for the narrowest necessary, and so do we.

activeTab over all sites

Chrome's activeTab page notes that an attacker who compromised such an extension would have to wait for the user to invoke it before getting access.

Logic on your server

Prices, keys and private data stay server-side, where you can rotate a credential without shipping an update.

Readable code

Google bans obfuscation and allows minification. Readable code is code a reviewer, and your next developer, can check.

Sensitive values in memory

Chrome's storage reference recommends session storage, held in memory while the extension is loaded, for sensitive user data.

Our own GatorCrawl Keywords extension is built on those rules. It requests storage and contextMenus and nothing broader, talks to one server, its own web app, and runs its content script only on Google search results. Its service worker accepts a sign-in token "only from the configured app origin," and its header comment reads "No analytics, no tracking, no browsing history." Its bar on Google renders inside a shadow root, so the page's styles and the extension's never mix. The build story shows the code behind each choice, and our permissions guide explains the warnings users see.

For the IT side: deciding which extensions staff can run

Businesses on managed Chrome have more control than most realize. Google's ExtensionSettings documentation lets an administrator block, allow, force-install or remove extensions, and goes further with fields such as "blocked_permissions", "runtime_blocked_hosts" (up to 100 entries as of November 2026) and "minimum_version_required". In plain terms, IT can refuse any extension that asks for a particular permission, keep extensions off sensitive sites such as the payroll system, and require a minimum version.

Force-installing is the other half. Google's force-install help page says users cannot remove force-installed items, and it is candid about the trade: "Force-installing an app or extension gives it permission to access information on the device it's installed on." That is a strong reason to force-install only tools whose code and publisher you trust, ideally your own. If the honest answer is that nobody in the company owns these decisions, a fractional CTO can.

What we promise, and what nobody should

We build extensions with short permission lists, logic on the server, readable code and a written data map, and we publish them under the client's own developer account so the publishing keys belong to the business. Store submission, the privacy disclosures and the permission plan are part of the build, from $1,500, depending on complexity, with a fixed written estimate before work starts; our pricing guide explains what moves that number. Ongoing updates are quoted, as our maintenance page explains. We do not promise an extension that cannot be compromised. Anyone who does is selling the promise.

The same thinking applies inside WordPress, as our page on plugin security shows. For the offer itself, see the extension build service; for the store's rules, the review guide; for everything else, the section overview. If an extension needs to connect to your CRM or internal systems, our API integration work goes live the same day. DataCram is a San Antonio web studio, building since 1999 (about); contact us or call (210) 346-0848.

Questions owners ask about custom browser extensions

According to Cyberhaven's own post of December 27, 2024, read through an Internet Archive copy, a phishing attack on December 24 compromised an employee's Chrome Web Store access, and the attacker published a malicious version, 24.10.4. Cyberhaven says the code was active from 1:32 AM UTC on December 25 to 2:50 AM UTC on December 26, affected Chrome-based browsers that auto-updated in that window, and was replaced by version 24.10.5.

It can reach whatever its permissions allow. Cyberhaven said the malicious version of its extension could have exfiltrated cookies and authenticated sessions for certain targeted websites. That is why the permission list matters: an extension limited to one site and two features has far less to give away than one with access to every website.

Treat it like a new vendor. Reread its permissions and privacy disclosures, look at what recent updates changed, and remove it if nobody can vouch for the new publisher. The Great Suspender's original developer announced in June 2020 that a new owner had bought the rights to publish it; from then on, updates came from that owner.

Open it on a company-controlled email, turn on 2-Step Verification with a security key, which Google's security guidance prefers, and never grant API access from a link in an email. A Chrome Web Store team post in January 2025 said the team will not ask developers to grant access to their account through its API.

Yes, on managed Chrome. Google's ExtensionSettings policy lets administrators block, allow, force-install or remove extensions, block specific permissions through blocked_permissions, keep extensions away from up to 100 hosts with runtime_blocked_hosts, and set a minimum version. Our internal team extensions page covers deploying your own tools the same way.

Not automatically. A custom extension is as safe as its code, its permission list and the account that publishes it. What custom gives you is control: you know who wrote it, what it requests, and who can publish updates, because the developer account is yours. Nobody can promise it will never be compromised.

Keep reading

Free · No obligation

Have someone careful look at your extension

Tell us which extension you build or rely on and what it touches. We reply with what we would check first, from permissions to the publishing account, and what a review or rebuild would cost.

Which of these sounds like you? (pick any)

Rather talk now? Call (210) 346-0848.

CallGet a quote