On November 16, 2026, researchers traced one OPTIMJACK delivery path through Optimizely project scripts; other affected pages used different script paths. StepSecurity reports that researcher Scott Helme counted 186 affected sites and 188 infected script files. Those are exposure counts, not a count of visitors, stolen cards, or compromised devices.
Report URI’s investigation found 188 infected snippets associated with 178 organisations in its own sample. The researchers examined 1,611 distinct Optimizely snippets they could find before the cleanup. These measurements overlap with Helme’s observations, so they cannot be added to form a larger total.
- 1,611Distinct Optimizely snippets examinedSource: Report URI, November 16, 2026
- 188Infected snippets foundSource: Report URI, November 16, 2026
- 178Organisations linked to infected snippetsSource: Report URI, November 16, 2026
What Do the OPTIMJACK Counts Actually Measure?
The unit matters because a site can load more than one script, and one organisation can run multiple sites. Report URI counted files in a bounded crawl, then linked the infected files to organisations. StepSecurity cited Helme’s separate site count and confirmed an Optimizely-hosted loading chain on example pages.
Neither study measured how many people loaded the altered code. Report URI says its sample could miss snippets it did not locate in time. Its 188 infected files therefore describe observed exposure in that sample, while the visitor impact requires different evidence.
Report URI’s sector table places 32 organisations in banking and finance, 23 in technology and software, and 20 in retail and fashion. These are categories within its 178-organisation set, not separate new incidents. The distribution points defenders toward pages where an outside script can touch payment or account workflows.
The five bars show selected sectors, not all 178 organisations. Sector labels and counts follow Report URI’s classification, and the bars do not establish that any visitor was harmed. A security team should map the finding to its own pages and traffic before estimating customer impact.
How Did a Trusted Script Become an Entry Point?
Optimizely Web lets customers load project JavaScript from its CDN. According to Report URI, malicious code appeared in the Project JavaScript portion of affected customer snippets. Direct CDN loads still arrived from Optimizely’s trusted host, while some customers served the same project content from their own domains. Neither path would put the change in the site’s own code repository.
Report URI observed several loader forms, including an encoded variant and an inline payload. This means searching current files for one literal string cannot reconstruct every earlier exposure. StepSecurity independently confirmed a chain in which a consent manager loaded an Optimizely file, which then loaded attacker-controlled code.
The researchers have not established how the attacker gained write access across customer projects. Report URI lists several possibilities but leaves the access route unresolved. Treat the project configuration and provider account audit as open questions while inspecting the browser behavior that is already documented.
This event illustrates the software supply chain compromise technique at the browser delivery layer. The useful boundary is each script’s actual behavior, not only its approved host. A third-party risk management platform can help keep ownership and provider review attached to that dependency.
What Could the Browser Code Reach?
StepSecurity’s payload analysis describes wallet address substitution, payment-card skimming, and conditional collection of screenshots and accessible cookies. The functions depended on what a visitor did on the page. The presence of an infected script alone does not prove a card or wallet was stolen.
Report URI says card collection watched site-controlled forms and outgoing requests. Payment details entered inside a cross-origin hosted payment iframe were outside that code’s reach. For each affected page, defenders should check whether payment fields were in the site’s own DOM or isolated in a provider frame.
The researchers also saw an iPhone and iPad branch loading additional code. StepSecurity did not retrieve or analyse that second stage, so it did not confirm its exploit chain or affected versions. Report URI described exploit kits in its analysis, but the actual outcome on individual devices remains unmeasured here.
Report URI linked observed wallet transactions to about $7,040 received as of 19:25 UTC on November 16. StepSecurity separately estimated about $6,600 from the wallets it checked and noted that some addresses were active before this event. Neither estimate establishes campaign proceeds or the number of victims.
What Should an Exposed Site Check First?
Start with a third-party script inventory for public pages, login, wallet, and checkout flows. Record each Optimizely project, the pages that load it, and the team with change authority. The third-party risk lens for supply chain security provides a broader ownership framework.
Preserve dated copies of script responses, response headers, and browser network logs before relying on a current clean fetch. Report URI found the infected snippet count fell from 169 at 20:56 UTC to one at 21:52 UTC on November 16. That cleanup means a clean file today cannot settle what visitors received earlier.
- Inventory scriptsMap Optimizely projects and other vendor files to the pages that load them.
- Preserve evidenceSave dated script responses, headers, CSP reports, and browser network records.
- Check page behaviorLook for unexpected child scripts and outbound connections in the affected window.
- Inspect paymentsConfirm whether card fields use isolated provider frames or site-controlled forms.
- Confirm cleanupAsk the provider for affected project IDs, change times, and the access review.
Review Content Security Policy reports for newly requested child scripts and unexpected outbound connections. An enforced policy with suitable script and connection restrictions may block part of this chain, according to both researchers. Test the actual policy because trust delegated to scripts can change the result.
Inspect payment and wallet journeys separately from general site traffic. If a site-controlled form accepted card details during the exposure window, follow the organisation’s payment incident process. For wallet addresses, compare the intended destination against a trusted out-of-band record, not copied page text.
Use third-party risk management records to assign the provider follow-up and track exceptions. Ask for a project-level change history and a timeline of file restoration. Keep the investigation open until the site can explain both the injected script path and its own customer exposure.
Method and Limits
This article compares two primary investigations published around November 16 and 17, 2026. Report URI’s counts come from its crawl of 1,611 distinct snippets; StepSecurity cites Helme’s 186-site count and documents its own sample checks. Both cover the same event, so their file and site figures are not summed.
The documented script behavior establishes a route to data in an affected browser, with different triggers for cards, wallets, and screenshots. It does not establish an affected visitor total, a complete loss figure, or a confirmed iOS second-stage outcome. The route into Optimizely project settings remains a question for the provider’s investigation.
Frequently Asked Questions
Does 186 mean 186 companies lost data?
No. StepSecurity cites 186 affected sites. The count does not measure visitors or confirmed data loss.
Why did Report URI count 188 snippets and 178 organisations?
It examined 1,611 distinct Optimizely snippets and found 188 infected files linked to 178 organisations. A company may use more than one snippet.
Can a clean Optimizely file now rule out earlier exposure?
No. Report URI documented rapid cleanup on November 16. Review saved responses, headers, CSP reports, and browser records for the exposure window.
