In short
Two publishers reported malicious code reaching users through trusted platforms. The Record reported Optimizely confirmed hackers inserted malicious code into customer sites, and Dark Reading reported Cursor traced a malicious Claude Code extension to a third-party breach. Both point to supply chain compromise rather than direct attacks on each victim.
What the headlines say
Two separate reports in the last 48 hours describe malicious code reaching users through trusted software channels. According to The Record's headline, website management platform Optimizely confirmed that hackers put malicious code into customer sites. Separately, Dark Reading reported that Cursor traced a malicious Claude Code extension to a third-party breach.
We have only the headlines here, not the underlying articles, so we are limited to what they state. Both items are tagged as breaches, and the Cursor report additionally touches on AI and identity. We cannot confirm scope, timelines, how many customers were affected, or the specific technical mechanism from the headlines alone.
Is this a cyber attack?
Based on the reporting, yes, both read as intrusions rather than accidental outages. The Record's wording that Optimizely "confirmed hackers put malicious code into customer sites" describes deliberate insertion of code by an outside party. Dark Reading's phrasing that Cursor "traced a malicious Claude Code extension to a third-party breach" points to a compromise of a supplier upstream of Cursor's users.
How can organizations tell an attack from a glitch? A malfunction tends to produce errors, downtime, or corrupted output with no sign of unauthorized access. An attack usually leaves evidence of intrusion: unexpected code changes, unfamiliar accounts or tokens, outbound connections to unknown destinations, or files modified outside normal deployment processes. When a vendor publicly uses the word "hackers" or "breach," as reported here, that is the clearest signal.
Why supply chain exposure matters
Both reports share a theme: the malicious code did not originate inside the victim organizations, it arrived through something they trusted. With Optimizely, the reported injection reached customer sites through the platform itself. With Cursor, the reported path ran through a third-party breach that produced a malicious extension.
That is the defining feature of a supply chain attack: one compromise can fan out to many downstream users who never interacted with the attacker directly. The Cursor case also illustrates excessive agency, where AI coding tools and their extensions can run with broad access to code, credentials, and systems, so a single poisoned extension can do outsized damage.
What organizations should check now
If you use either platform, treat vendor advisories as the authoritative source and act on them as they are published. In the meantime, several steps are useful generally:
- Inventory where these tools touch your environment, including embedded scripts, browser or editor extensions, and CI/CD integrations.
- Review recently added or updated third-party code and extensions for anything you did not approve.
- Rotate credentials, API keys, and access tokens that the affected platforms could read or use.
- Pull logs for the suspected exposure window and look for unusual logins, new sessions, or outbound traffic to unfamiliar hosts.
- Verify the integrity of deployed site code against a known-good baseline.
For AI developer tooling specifically, limit what extensions can access, require review before installing new ones, and separate secrets from the environments those tools run in.
How this connects to our guidance
These incidents are a reminder that trusted platforms are part of your attack surface. Our Threat Radar tracks emerging exposure, and the Field Manual breaks down techniques like the two patterns above. Our research and briefing pages offer deeper context as stories develop, and if a compromise is confirmed in your environment, incident response planning should kick in quickly.
StrikeCyber helps organizations test how far an injected extension or poisoned dependency could reach through realistic penetration testing.
Sources
- Website management platform Optimizely confirms hackers put malicious code into customer sites, The Record, 2026-11-17
- Cursor Traces Malicious Claude Code Extension to Third-Party Breach, Dark Reading, 2026-11-17
Frequently asked questions
Were these confirmed cyber attacks?
The Record's headline says Optimizely confirmed hackers put malicious code into customer sites, which indicates an intrusion. Dark Reading reported Cursor traced a malicious Claude Code extension to a third-party breach, which also points to a compromise. Both are described as breaches in the headline topics, so these read as attacks rather than accidental faults.
What is a supply chain attack?
A supply chain attack compromises a trusted vendor, platform, or component so that malicious code reaches many downstream customers at once. Instead of attacking each organization directly, the attacker poisons something many organizations already trust and run. These two reports both fit that pattern, affecting website customers and extension users respectively.
What should my organization do right now?
If you use Optimizely or Cursor, watch official vendor advisories and apply any guidance as it is published. Review recently loaded third-party scripts, extensions, and dependencies for unexpected changes. Rotate credentials and tokens that those platforms could access, and check logs for unusual activity during the exposure window.
