Fake OpenAI/SpaceX Event Invites Beat MFA on Work Accounts

The invitation looks like a perk. A November 4 OpenAI livestream, "ChatGPT for Social & Digital Marketing Teams," with three months of ChatGPT Pro for attendees. Or a SpaceX "Starship Product Briefing" on November 12, offering three months of Grok Heavy or a "SpaceX share." All you have to do is sign in with your work Microsoft or Google account.

Neither event exists. Push Security published its analysis on November 5, 2026, after blocking visits to the fake pages in late October: "Employees want access to the latest AI tools, making them excellent bait for attacks."

Key Takeaways

• Push Security reported on November 5, 2026 that fake OpenAI and SpaceX event pages ask visitors to sign in with a work account to claim AI subscriptions.
• Kali365 runs device code phishing against Microsoft accounts, while VoidProxy proxies real Microsoft and Google logins to capture passwords, MFA responses and session cookies.
• Eventrsvp[.]app, the lure domain behind both pages, was three days old on November 2, and the operator rotates domains every few weeks.
• Push published no victim counts, so the campaign's scale is unknown.
• A passkey defeats the VoidProxy branch, but the Kali365 branch completes on Microsoft's real sign in page, so only refusing device codes from websites or blocking device code flow stops it.

What Is the Fake OpenAI and SpaceX Event Campaign?

It is a session theft campaign that dresses phishing pages up as AI brand events. The Push Security analysis describes two live variants copying the look of the OpenAI Academy events site and a SpaceX AI for marketing briefing. A third, "Anthropic at AWS re:Invent," had already been taken down, according to BleepingComputer.

Two page kits build the lures: ZoomMock, which made fake Zoom event pages from May to October 2026, and MSDRoot, first seen on October 22. They share no code, but Push links them with "medium-high" confidence through a shared domain and the same targeting: "Both kits target marketing staff with AI-brand events."

How Do the Invitations Reach Victims?

They arrive by email, routed through link services that mail filters tend to trust. Push lists Google short links (c.gle and share.google), Zapier's zpr.io and an x.com redirect, calling them "common techniques for evading reputation-based URL checks." A Cloudflare challenge then hides the lure page from most public scanners.

Earlier waves leaned on real platforms. Cofense reported on October 29 that invitations from "Austin from nthropic" (the A was missing) came through GoTo Webinar's customercare@gotowebinar[.]com sender. Others linked to genuine events.zoom.us pages. It matches what KnowBe4 found about phishing sent through legitimate services: the sender passes every check because it is real.

How Does the Attack Get Past MFA?

It steals what MFA produces, OAuth tokens or a signed in session cookie, instead of guessing a code. Push describes Kali365 and VoidProxy as "separate phishing-as-a-service (PhaaS) platforms, both designed to bypass or defeat MFA checks." Each event page offers two buttons.

• Continue with Microsoft. The page fetches a device code from Kali365, shows it in a "Confirm access" dialog, and opens Microsoft's genuine device login page in a popup. When the victim pastes the code and approves, Microsoft issues tokens to the attacker's Kali365 session. The page then forwards the victim to a VoidProxy Microsoft login that grabs the password and session cookie as well.
• Continue with Google. A popup titled "Portal" opens on a VoidProxy reverse proxy domain that relays Google's real sign in and captures the password, the MFA response and the session cookie.

The FBI's May 21 alert on Kali365 said the kit was "first seen in April 2026" and "has primarily been distributed via Telegram" (see our Kali365 explainer and our September VoidProxy breakdown). Push found no tie between the two developers and believes the operator is "most likely a customer of both." Six months after Kali365 appeared, its customers are chaining it with a rival kit.

Why Does a Passkey Only Stop Half of This?

A passkey stops the VoidProxy branch but not the device code branch, because device code phishing happens on Microsoft's real website. Phishing resistant MFA such as FIDO2 keys works against the proxy pages: a passkey is bound to the genuine domain and will not sign in on a lookalike, which is why Google's passkey page promises "stronger protection against threats like phishing."

On the Kali365 path, the victim signs in on Microsoft's own device login page, so a passkey works as designed and still approves the attacker's session. Only the code's origin is fake. Microsoft documented the same trick in its Storm-2372 report, a campaign active since August 2024 that used fake Microsoft Teams meeting invitations delivered through email. A device code fits naturally into a join step.

Which Domains Should Defenders Watch?

Push lists these indicators for the October and November activity, with a warning that "The operator rotates domains every few weeks, so these indicators have limited value on their own." The Push report has the full table.

• Lure pages: openai.eventrsvp[.]app, spacex.eventrsvp[.]app, anthropic.rsvpconnect-events[.]com
• Collection server: 217.60.102[.]48
• Kali365 backend: 62zr4vcasl.shifts[.]icu
• VoidProxy Google proxy: claudewebinnar[.]com and its subdomains
• Microsoft sign in proxies: wbdssologin[.]app, continue-to-office-.ggleventapi[.]icu

Why Email Users Should Care

Everything before the fake sign in page exists to get an invitation past your filters: a real GoTo Webinar sender, a real Zoom page, a Google short link. Last month brought the same playbook with fake calendar invites and fake YouTube sponsorship pitches. The bait is an opportunity; the ask is a sign in.

The prize is the mailbox. Tokens or a session cookie for Microsoft 365 or Google Workspace open email and files without another MFA prompt. The pages say a work account is required, and Push does not say whether personal Gmail accounts work, so the work account is the target. Its vendor contacts will trust its next message.

What Should You Do Right Now?

1. Verify the event yourself. Type the company's address into your browser and find the event there. A freebie that needs your work login is the red flag.
2. Never enter a device code a website gives you. A legitimate code appears on a TV or command line you started yourself, never in a popup on an event page.
3. Check your sessions. In Google, open Security & sign-in, then Manage all devices, and sign out of anything unfamiliar (Google's steps). For a Microsoft work account, review Recent Activity at myaccount.microsoft.com, change your password if anything looks wrong, and ask IT to revoke your sessions.
4. Add a passkey. Create one at myaccount.google.com/signinoptions/passkeys to shut down the proxy branch.
5. Admins: block device code flow. Microsoft's Conditional Access guidance says "We recommend organizations get as close as possible to a unilateral block on device code flow." Audit existing use first.