MediumSecurity incidentupdated 29 Oct 2026, 21:56 UTC

Fake Anthropic-themed emails are slipping past email security tools

Scam emails pretending to come from Anthropic, the AI company, are getting into inboxes even though they have obvious warning signs. Those emails can trick people into clicking links or handing over information.

Attacks reported (Dark Reading)1 source, unconfirmed

Severity	MediumSecurity incident
Being exploited?	ReportedBy Dark Reading, not officially confirmed yet
What to do	Check exposureWarn staff now that Anthropic-themed emails with urgent requests may be phishing, even if they pass your email filter.
How sure is this?	1 source, unconfirmedOnly one publisher so far, details may change

What happened

Phishing emails impersonating Anthropic are bypassing several leading secure email gateway (SEG) products despite containing numerous red flags.

Why it is dangerous: Employees and users who rely on email filtering may receive malicious messages that look trusted, increasing the chance of credential theft or malware infection. Organizations using affected SEG products are most exposed.

Are you affected?

You are affected if your organization uses one of the secure email gateway products that are letting these phishing emails through and you receive mail claiming to be from Anthropic.

What to do

Warn staff that emails claiming to be from Anthropic may be fake, especially if they ask for logins, payments, or urgent action.
Tell staff to check the sender address and hover over links before clicking.
Ask your email security vendor whether these messages are being flagged and whether additional rules can block them.

Timeline

29 Oct 2026, 21:21 UTC
Publicly disclosed
29 Oct 2026, 21:21 UTC
Reported by Dark Reading · Dark Reading
29 Oct 2026, 21:56 UTC
Picked up by Zero Radar

Sources 1 report, oldest first

Security press established security newsrooms

Anthropic-Themed Phishing Emails Bypass Top Security ToolsDark Reading · 29 Oct 2026, 21:21 UTC

Technical details

ID: ZR-cbcc5cba
Disclosed: 29 Oct 2026, 21:21 UTC
On Zero Radar: 29 Oct 2026, 21:56 UTC (35 min after disclosure)
Last change: 29 Oct 2026, 21:56 UTC

The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works