Loading
Bolingbrook, Illinois · Bilingual EN/ES team · SAM.gov registered +1 646 687 6300

One script infected 178 sites: BBC, CNN, Zoom and 32 in banking

On November 16, 2026, a flaw in Optimizely let attackers inject malicious code into the websites of 178 organizations, 32 of them in banking and finance. Opening the page was enough. Here is what happened and what to check on your own website today.

Un script de terceros comprometido en el centro de una red que distribuye código malicioso a decenas de ventanas de sitios web

On Monday, November 16, 2026, anyone who opened the BBC, CNN or Zoom could be served malicious code without clicking anything. Nobody hacked those sites: attackers hacked a component they all loaded. Here is what happened and, above all, what you should review, because your website also runs code written by another company.

What happened with Optimizely on November 16

The discovery came from researcher Scott Helme and his team at Report URI. Checking 1,611 Optimizely snippets, they found 188 infected ones across 178 organizations, including Zoom, BBC, CNN, Salesforce, GitLab, IKEA, HSBC and Equifax. Banking and finance was the hardest-hit sector, with 32 organizations. Report URI warns the real number is likely higher.

The attacker never touched those companies' servers. The code was written into a field in each Optimizely project that runs on every page, so the script kept arriving from the vendor's legitimate domain.

  1. 04:21 UTC

    First injection

    Report URI sees the first altered snippet, the BBC's.

  2. 08:28 UTC

    Crypto theft

    The code that replaces wallet addresses starts loading.

  3. 13:43 UTC

    Card skimming

    The attacker adds a skimmer that captures number, expiry and CVV.

  4. 19:24 UTC

    Discovery

    Report URI's sweep identifies the infected snippets.

  5. 23:11 UTC

    Cleanup

    Every file Report URI was monitoring shows as clean.

Un formulario de pago del que se desvían datos por un canal oculto, junto a otro formulario aislado en una cápsula que permanece protegido
Cards typed into a site's own forms were within the skimmer's reach; cards entered in fields hosted by the payment processor were not.

Optimizely confirmed that an unauthorized party exploited a vulnerability in its experimentation product to add malicious JavaScript to some customer projects. It says it blocked the vulnerable path, removed the code, ended sessions and revoked tokens, and is investigating with outside experts. It has not explained how the flaw was exploited.

Report URI traced about $7,000 in cryptocurrency sent to the attacker's wallets. What remains unknown is how many cards, cookies or devices were compromised.

Why does this matter to a mid-sized company?

Because the problem is not Optimizely. It is the model. Analytics, chat, advertising, A/B testing: every third-party script has, inside your page, the same permissions as your own code. And the attack asked nothing of the visitor; according to researchers quoted by KrebsOnSecurity, the code ran as soon as the page loaded.

The good news is in the same data: some sites loaded the infected script and nothing happened, because their content security policy (CSP) blocked it.

68affected sites with no CSP at all
50with a CSP that still allowed the code and the data theft
24with a CSP that blocked the malicious code

Report URI, analysis of affected sites' policies, November 16-17, 2026

Trusted JavaScript doesn't always remain trustworthy.

Scott Helme, Report URI, quoted by KrebsOnSecurity

Four steps for your website

Inventory your scripts

List every third-party script your site loads, who asked for it and why. Remove the ones nobody uses anymore.

Turn on a strict CSP

Limit which domains can run code and where the page can send data. Avoid 'unsafe-eval' and enable reporting.

Isolate payments

Use your payment processor's hosted fields instead of your own form fields for card data.

If you use Optimizely

Review your projects' JavaScript, rotate credentials and tokens, and treat cards typed into your own forms on November 16 as exposed.

Una compuerta de filtros que deja pasar los scripts autorizados y detiene los maliciosos antes de que lleguen al sitio web
A content security policy decides which code may run on your page and which may not.

At Hitos we see this case as an uncomfortable reminder: your site's security also depends on vendors you do not control, and the first step is knowing what you have exposed.

Sources: Report URI · The Record · KrebsOnSecurity · Malwarebytes

Source: Report URI

Keep reading

Security

The era of the AI cyberattack has begun: what happened and what to do

In four days, the FBI warned about AI agents being used to attack companies, Google documented an attack run with no humans involved, and California sued OpenAI over the agents that attacked Hugging Face. Here is what happened and what your company can do.

· 2 min read

Have a system to connect or a process to automate?

Talk to us