Channel sheet · CH-16 · gain 4 min · logged November 18, 2026
Martech StackDirect input
Optimizely Signup Flaw Fed Malware to 178 Organizations' Sites
A signup-flow flaw in Optimizely let an attacker inject wallet-swapping, card-skimming JavaScript into 178 organizations' projects. Only 24 sites blocked it outright; attacker wallets took in about $7,040.
By Sophie Lindqvist4 min read880 words
Signal notes
- Malicious code found in snippets of 178 organizations, including Zoom, BBC, CNN, Salesforce and GitLab, on Nov. 16.
- Attacker wallets received approximately $7,040 by 7:25 p.m. UTC on Nov. 16.
- Only 24 of 159 reviewed organizations (15%) fully blocked the malicious code via restrictive script-src policies.
- 188 of 1,611 reviewed snippet files contained malicious code; files were clean by 11:11 p.m. UTC on Nov. 16.
- Optimizely said it blocked the vulnerable signup-flow path, deployed a fix, and reset access for affected accounts.

Malicious JavaScript reached the websites of 178 organizations — including Zoom, the BBC, CNN, Salesforce and GitLab — on Nov. 16 after an attacker exploited a flaw in Optimizely's account signup flow. The attacker's wallets collected roughly $7,040 by 7:25 p.m. UTC that day, and only 24 of the studied sites blocked the code from running at all.
Scott Helme, founder of Report URI, a company that sells tools for monitoring the scripts websites run, reported the findings. Optimizely's status page acknowledged that an unauthorized party "exploited a vulnerability in Experimentation" to add malicious JavaScript to some customer projects. In a statement to BleepingComputer, the company said it has "blocked access to the vulnerable path and deployed a fix" and has "ended active sessions, revoked tokens, and reset access for affected user accounts."
How did the code reach customer sites?
Optimizely Web works by having each client load a JavaScript snippet hosted on Optimizely's content delivery network. Besides experiment data, the snippet carries an optional Project JavaScript section where clients can put their own code that runs on every page load.
The attacker placed malicious code in that section and left the rest of the script intact. The script still came from cdn.optimizely.com — a domain every affected site already trusted. Scripts clients served from their own domains were affected too, along with those for optimizely.com and login.optimizely.com.
What did the injected code do?
The inserted code loaded a second script with several jobs:
- Replace cryptocurrency wallet addresses with the attacker's.
- Starting at 1:43 p.m. UTC, watch the site's own form fields for card number, expiration date and CVV, and submit all three to the attacker once captured.
- When swapping an address near terms like "wallet" or "payment," also send the page URL, cookies not marked HttpOnly, and a screenshot to the attacker's server.
- Load the Coruna and DarkSword exploit kits against iPhones and iPads running older iOS versions.
Cards entered through hosted payment iframes such as Stripe and Adyen stayed out of reach. Report URI cannot say how many cards, cookies or devices were compromised.
Helme's team reviewed 1,611 Optimizely snippet files, mostly from Report URI's own crawl data; 188 contained malicious code. Helme believes the true count is higher. The code first appeared at 4:21 a.m. UTC on Nov. 16, and all infected files were clean by 11:11 p.m. UTC the same day.
Why didn't Content Security Policies stop it?
A Content Security Policy tells the browser which scripts a site may run (script-src) and where it may send data (connect-src). Report URI compared policies across 159 affected organizations:
- 68 had no Content Security Policy at all.
- 50 had a policy that permitted both the malicious code and the outbound transmission of user data.
- 17 allowed the code to run but blocked data exfiltration through connect-src limits — though wallet address swaps still succeeded.
- 24, or 15%, blocked execution entirely; they shared a restrictive script-src directive.
Helme wrote that allowlisting Optimizely's domain offers no protection here, because the malicious code originated from that same domain. The common 'strict-dynamic' setting grants access to code inserted via the Optimizely snippet, and 'unsafe-eval' allows an inline version to run even when the original is denied. Optimizely's own CSP help page recommends 'strict-dynamic' and includes 'unsafe-eval' in its sample policy, since features like audiences with custom JavaScript conditions require it.
What should teams check now?
Helme says teams should open Project JavaScript in all their Optimizely Web projects and the custom JavaScript in experiments, looking for 1688360.cc, atob( and any scripts not written by their own developers. They should also review who holds write access among users, API tokens and integrations, then rotate credentials. Sites with card fields in their own forms should treat every card entered on Nov. 16 as possibly exposed, and rotate any session token or secret stored in a cookie not marked HttpOnly.
Optimizely's change history documentation says each project's History logs every Project JavaScript change, filterable by how the change was made — via the REST API or the Web Experimentation interface. The status page lists no customer steps.
What does this change for operators?
Project JavaScript runs on every page that loads an Optimizely snippet, so the attack hit visitors without touching the affected sites' own code. That puts the Optimizely account — who can edit Project JavaScript, which API tokens exist — on the same audit list as the site itself.
A vendor domain on an allowlist did nothing against code served from that domain. The 24 organizations that blocked the code had script-src lists limited to only the hosts they needed. Where the code still ran, connect-src limits kept card, cookie and screenshot data from leaving.
Optimizely told BleepingComputer it continues to investigate and will post updates on its status page, which listed the incident as Monitoring on the morning of Nov. 17. The company has not said how many customer projects were affected, or how a signup-flow flaw let the attacker modify existing projects. Until it does, the checks inside the account are the ones teams can run themselves.
via searchenginejournal.com (Source)
More from Sophie Lindqvist
Show full bio
Staff writer covering money, funding and market shifts around marketing professionals running AI in the stack at Mart Signal.
169 articles
Bus out
- Google Deploys AI Spam Filter Called SAFE, Search Engine Journal Reports
- Safari's Identity Block: The Tide Just Went Out on Ad Tech
- 194 AI Visibility Tools in Two Years, Most Will Not Survive
- Top Million Website Visits Remain 12.2% Below 2022 Peak
- Dentsu ships AI-powered influencer marketing operating system