Fake YouTube sponsorship emails are hijacking creator channels: the info-stealer behind the "brand deal"
A flattering sponsorship offer from a real-looking brand lands in a creator's inbox. To "see the deal" they download the media kit, which quietly installs an info-stealer that copies their browser session cookies. Hours later the attacker is signed into the channel with no password prompt, has locked the creator out, and is live-streaming a fake crypto giveaway to their subscribers.
Bottom line first
A YouTube sponsorship email that pushes you to download a media kit, contract, or brief before you can even see the terms is almost always malware, not a real brand offer. Opening the file, often a password-protected .rar, .zip, or .exe, installs an info-stealer such as RedLine, Vidar, or Lumma that copies your browser session cookies and saved passwords. With a stolen YouTube session cookie the attacker signs straight into your Google account without ever needing your password or your two-factor code, because they reuse the session you are already logged into. They lock you out, rebrand the channel to impersonate a crypto company, and run live "giveaway" streams to your own subscribers. Real brands reach out from their verified corporate domain, for example nvidia.com, adobe.com, or epicgames.com, and never force a download to view a brief. A second version of the same campaign, documented by ESET in October 2026 but running since at least June 2026, asks for no download at all. It walks you through a polished "creator sponsorship platform" and then asks you to confirm channel ownership by signing in with Google, except it never sends you to Google: it draws a counterfeit browser window inside the page, complete with its own address bar reading accounts.google.com and a padlock icon, and puts the sign-in step in that frame. Your two-factor code is genuine and correct, which is exactly why nothing feels wrong, and it is captured next to your password and used to sign in as you. So the absence of an attachment is not reassurance, and neither is an address bar you can see inside the page, because the page drew it. No real brand makes you download a file to read a brief, and none needs you to prove you own your channel by signing into Google on their site. Never sign in from a link in an offer: open a new tab, type the address yourself, and run the platform link through a scam checker before you click anything in the email. Do not open any attachment or link from an unsolicited sponsorship email, and verify the offer by going to the brand's official site yourself.
How one sponsorship email ends with a hijacked channel
The whole attack is a chain, and every link looks harmless on its own. A creator gets an email that reads like a genuine partnership pitch from a household brand. It offers real money for a dedicated video or an integration, and it attaches or links a "sponsorship brief" and "media kit" so the deal feels concrete. The creator, excited, opens the file. That single click installs an information stealer. From there the attacker does not fight your defenses, they walk around them: the malware exfiltrates the browser session cookie that keeps you logged into YouTube, the attacker loads that cookie into their own browser, and your account opens for them as if they had typed the password themselves. They change the recovery email and phone, remove you, rename the channel to look like a crypto exchange, and go live with a scam giveaway aimed at the audience you spent years building.
This is trending right now. In 2026, security teams at Bitdefender ("YouTube warns creators of targeted AI-powered phishing"), Malwarebytes, and McAfee all documented this wave of creator-account phishing, and Google issued a fraud advisory in June 2026 warning creators about impersonated brand outreach. Multiple creators have reported live channel takeovers within the same week. The bait is old, but the delivery is polished enough to fool people who know better.
The lure: a paid brand deal in your creator inbox
The email is written to flatter and to move fast. A named "partnerships manager" from a brand you would love to work with says they have been watching your channel, that your audience is a perfect fit, and that they are ready to pay for a sponsored video. The numbers are attractive but not absurd. The tone is professional. There may be a logo, a signature block, and links that look like the brand's real site at a glance.
Then comes the ask that matters: to move forward, download the attached brief, media kit, or contract. Sometimes it is a link to a file on a cloud host. Sometimes it is a password-protected .rar or .zip (the password "protects" the file from your antivirus, not from you). Occasionally it is a direct .exe dressed up as a PDF viewer or a "campaign dashboard." Creators are used to receiving brand assets, so a media kit does not feel out of place. That familiarity is the whole point. It is the same psychology behind the fake YouTube copyright-strike email and the fake YouTube Premium membership scam: dress the trap in the everyday paperwork of running a channel.
The payload: a "media kit" that installs an info-stealer
The download is not a document. It is a dropper for an information stealer, most often one of three families that dominate this kind of campaign:
- RedLine. A widely sold stealer that grabs saved browser passwords, autofill data, cookies, and cryptocurrency wallet files.
- Vidar. A modular stealer that pulls browser data, session cookies, and files matching common wallet and document patterns, then ships them to the operator.
- Lumma. A stealer-as-a-service popular in creator-targeting campaigns, built to harvest session tokens and browser-stored secrets quickly and quietly.
Whichever family it is, the goal is identical: read everything your browser has stored. That means saved passwords, autofilled logins, any browser wallet extensions, and, most valuable of all, your active session cookies. The malware often runs without any obvious symptom. There is no ransom note and no popup. The "brief" may even open a decoy document so nothing seems wrong. The same class of stealer is delivered through other creator-and-professional lures too, such as a fake meeting-link that drops malware when you "install the plugin" to join a call.
Catch a fake brand deal before it costs you the channel
Add the SafeBrowz browser extension, or the Android app, and it runs every check in this article automatically, on every page, before it renders. Free forever.
Need to check a suspicious link right now? Click here to scan it free →
Upgrade to Premium for AI deep scan that recognizes fake login and stealer-delivery pages in 100+ languages.
How the stolen cookie walks straight past your 2FA
This is the part that surprises careful creators: two-factor authentication does not stop it. Two-factor protects the moment you log in. But you are already logged in, and your browser proves that with a session cookie, a small token that says "this person already passed the login and the 2FA check, let them in." The stealer copies that token. The attacker imports it into their own browser and lands inside your active session. No password prompt appears. No 2FA code is requested. There is nothing for your authenticator app or your security key to challenge, because from Google's point of view the session already cleared every check.
Session-cookie theft is the same core weakness behind adversary-in-the-middle attacks that bypass two-factor, and it is why "I have 2FA on" is not a complete answer. Once the attacker is in, they move fast: change the recovery email and phone number, add themselves as a manager, and remove your access so you cannot simply log back in and undo it. Then they repurpose the channel. The usual play is to rename it after a well-known crypto brand, delete or hide your videos, and start a livestream promoting a fake giveaway ("send 1 coin, get 2 back") aimed squarely at your subscribers, who see it coming from a channel they already trust.
The other version: no download, a counterfeit sign-in window instead
Everything above describes the version that needs you to open a file. There is a second version that sends no file at all, and it has been running for months. On 7 October 2026, ESET's Christian Ali Bravo documented a brand deal scam targeting YouTube creators that impersonates Hollyland, which ESET describes as "a legitimate manufacturer of wireless transmission and audiovisual equipment", alongside what its own figure caption calls "virtually identical variants, posing as Nike and Spotify". ESET notes the domains and names changed repeatedly between June and August 2026, so this is not a new October campaign, it is a long-running one that was written up in October. The email is the same flattery, personalised with references to your actual videos. What it links to is not an attachment. It is a convincing collaboration platform, and the only thing it ever asks you to do is confirm you own your channel by signing in with Google.
ESET names two of the platform domains, joinmatchy[.]com and matchyjoin[.]com, and describes a variant it dubs "Scouty" on further domains. We found joinscouty[.]com, registered on 12 June 2026, sitting behind the same registrar and the same nameserver provider as those two, which is corroborating rather than conclusive: it points at one operator without proving it. A fourth was still serving the platform when we checked on 14 October 2026: mytubive[.]com, registered on 3 October, eleven days earlier. It brands itself Tubive, "The Global Infrastructure for Creator Sponsorships".
That live one is worth describing precisely, because the polish is the attack. It queries YouTube's public data API to pull your real channel name, avatar, subscriber count, video count and join date, so the dashboard greets you with your own numbers. It lines up a wall of brand logos, Hollyland, Nike, Samsung, Spotify, Discord, YouTube and NordVPN, and invents its own credibility: 14.2 million dollars "Paid to Creators", a "98.2% Escrow Success" rate. Then it offers to unlock your campaign brief and a Hollyland Lark M2 kit, a microphone that genuinely exists on Hollyland's own product pages, the moment you complete what it labels "Google Channel Verification".
We read that page's own scripts, and the sign-in is the part worth understanding. It does not send you to Google. It draws a browser window on top of the page. The script carries four encoded window skins, one for each combination of macOS or Windows and dark or light mode, plus two small behaviour scripts that make the result draggable. Every skin has a title bar, an address bar reading accounts.google.com, a padlock drawn as an inline graphic, and a frame for the content; the macOS pair goes further and reproduces the three traffic-light buttons, while the Windows pair wears Windows chrome with a Google logo in the title bar instead. Inside that frame the kit loads a page from its own host rather than from Google. We could not retrieve what that page serves, so for the content of the login step we are relying on ESET, which describes an imposter login page that captures the password and the one-time code. ESET's write-up of the earlier domains describes the victim being taken to a Google sign-in page, without saying whether that was a real redirect or a window drawn like this one, and it notes that the variants share portions of their source code. So we cannot tell from the outside whether the delivery itself changed between June and October or only the description of it did. This technique has a name and we have written about it already: it is a Browser-in-the-Browser attack, and the sponsorship platform is simply a better pretext than anyone had used for it before. ESET puts the outcome plainly: those credentials go, "and with them the account itself".
So your second factor is not bypassed by some clever flaw. You open your authenticator, read a genuine six-digit code, and type it into a form that is not Google's. The code is real and correct, which is exactly why nothing feels wrong, and it is captured next to your password and used to sign in as you. Separately, the script reports your progress back to its operator through a relay into a private messaging group: your IP address, the page you are on, and the visible text of whatever is in that frame, sent each time you move a step forward. It also reads a referral code out of browser storage, which is consistent with a kit rented out to more than one operator, though we only saw the code read, never sent.
This is where the usual advice runs out, and we should be straight about it, because the obvious instruction is to check the address bar and that instruction is not enough here. There are now two address bars on your screen: your browser's real one, at the very top of the window, and a picture of one that the page has drawn inside itself. The drawn one says accounts.google.com, with a padlock. Do not reach for the scroll test, by the way: this window is pinned in place and stays put when you scroll, exactly like real browser chrome. Three things do separate them. You cannot drag the drawn window out of the browser, because it lives inside the page. Its padlock is a static image, so clicking it does nothing, where your real browser's padlock opens site and certificate information. And its address is only text: a single click selects nothing, where one click in your real address bar highlights the whole URL ready to copy. The safe habit is simpler than any of that: do not sign in from a link at all. Open a new tab, type the address yourself, and sign in there.
Two things are genuinely reliable. Real Google sign-in happens on accounts.google.com, and the main legitimate exception is a channel managed through a company's Google Workspace, which can be sent to that company's own sign-in provider. Google's one-tap prompts are also real, and they appear over a partner's site with the partner's domain in the address bar, which is visually close to what this kit imitates, but they never ask you for a password or a code. If a password or a code is being requested, the only correct host is Google's own. And no genuine brand needs you to prove you own your channel by signing into Google on their website, because you replied from the address on it.
One last thing, because it tells you how long this will last. The three older domains have stopped serving the platform, but they have not gone away: two of them now answer with a shop-builder placeholder page, and all three are still registered into 2027. That is recycling, not eradication, and the newest one went up eleven days before we wrote this. Any advice built on recognising a particular bad domain expires the moment the operator registers the next one. Advice built on the shape of the request, a stranger asking you to prove you own your channel by signing in somewhere that is not Google, does not expire.
What the fake sponsor domains and files look like
The links and senders are built to look like a brand's partnership team at a glance. Because these lookalike patterns are illustrative and not tracked brands, the examples below are defanged and are not clickable. The shape is what to memorize, not the exact string:
- Brand-plus-offers on a domain the brand does not own: nvidia-brand-offers[.]com, adobe-creators-deal[.]com
- Sponsorship and media wording: epicgames-sponsorship[.]media-kit[.]co, brand-partners[.]creator-deal[.]site
- The download itself: a file named like
Sponsorship_Brief_2026.pdf.exeor a password-protectedMedia_Kit.rarwhere the password sits right in the email body
The habit that protects you is to ignore the brand name inside the address and read the actual host, the part right before the first single slash after the protocol. In nvidia-brand-offers[.]com the real host is nvidia-brand-offers[.]com, which is not NVIDIA. NVIDIA's real domain is nvidia.com, Adobe's is adobe.com, and Epic Games' is epicgames.com. Seeing a real brand name in a message is not proof of anything: verify by going to the brand's official site directly, never by clicking the link in the email. And when it is time to log in, YouTube sign-in happens at accounts.google.com, or, for a channel managed through a company's Google Workspace, at that company's own sign-in provider. It does not happen inside a page you reached from a sponsorship email. A convincing fake login screen served from any other domain is a phishing trick, the same Browser-in-the-Browser fake login window we break down separately.
The 2026 twist: an AI-generated "YouTube CEO" video
A newer variant adds a layer of manufactured authority. Instead of, or alongside, the sponsorship email, creators receive a private video or a link to one that appears to show YouTube's CEO announcing a policy change, a new monetization program, or updated "terms you must accept." The video is an AI-generated deepfake. Its only job is to make the follow-up instruction, download this file or sign in here to keep monetizing, feel official and urgent. It is a strong reminder that a familiar face or voice is no longer proof of anything. We track this and related tactics in our complete guide to AI-powered scams in 2026. The defense does not change because the packaging got smarter: YouTube does not deliver policy changes as a video that asks you to download a file, and it does not route you to a sign-in page through a link in an unsolicited message. Google's own sign-in lives at accounts.google.com.
Red flags in a fake sponsorship email
- You must download a file to "see the deal." Real briefs and rate discussions happen in the email body, a shared doc you can view in the browser, or a call. A required download, especially a .rar, .zip, or .exe, is the trap.
- A password-protected archive with the password in the email. That combination exists to slip the file past your antivirus, not to keep it secure. Legitimate brands do not send you password-locked executables.
- The sender domain is not the brand's real domain. A message "from NVIDIA" that routes through nvidia-brand-offers[.]com is not from NVIDIA. Real outreach comes from the brand's verified corporate domain.
- Urgency and flattery stacked together. "We love your channel, contract attached, we need an answer today" is a pressure pattern designed to skip your caution.
- A "verify your channel ownership" step on a platform you have never heard of. No genuine brand needs you to prove you own your channel by signing into Google on their website. They can see your channel is yours because you replied from the address on it. Treat that step as the attack itself, and judge it by the host in your real browser's address bar, not by any address shown inside the page. On this campaign that distinction is the whole ballgame, for the reason set out above.
- A sign-in prompt on a non-Google page. Any request to "log in with your YouTube account" on a page you reached from the email is credential theft. Real Google sign-in is at accounts.google.com, and the page that asks for it is never one a stranger's offer sent you to.
- A "policy" or "CEO" video that leads to a download or login. YouTube communicates through studio.youtube.com and official channels, not through a video that asks you to install something.
- The offer is too clean. No prior relationship, no negotiation, a generous flat fee, and an attached contract in the very first message is how these lures move you toward the file fast.
How to verify a brand sponsorship offer safely
Treat every unsolicited sponsorship email as a claim to check, not an instruction to follow. Confirm through a path you open yourself.
- Do not open the attachment or click the link. Not even out of curiosity. The download is the payload.
- Find the brand's real partnerships channel yourself. Type the brand's official domain directly, for example nvidia.com or adobe.com, and use the contact or partnerships page listed there to ask whether the outreach is genuine.
- Scan the link before you touch the file. Paste the landing page or download URL into the checker above, or into the full SafeBrowz scanner, and get a verdict first.
- Never sign in from a page you reached through the email. If something claims you must log in, open a new tab, go to studio.youtube.com yourself, and check your account there instead.
- Open anything unavoidable in isolation. If you genuinely must inspect a file, use a sandbox or a throwaway device that is not signed into your accounts, never your main creator machine.
If you already opened the brief
Assume the stealer already ran and that your session cookies and saved passwords are compromised. Move quickly.
- Get off the infected device. From a different, trusted device, go to accounts.google.com and change your Google password to a long, unique one you do not reuse.
- Kill every active session. In your Google security settings, sign out of all other sessions and devices. This is the step that actually invalidates the stolen cookie, so do it right after the password change.
- Review recovery and access. Check that the recovery email and phone are still yours, remove any unfamiliar ones, and review channel permissions and managers in studio.youtube.com. Revoke connected apps you do not recognize.
- Clean the machine. Run a full malware scan, and if the channel matters to your income, treat the device as compromised and rebuild it rather than trusting a single scan.
- Report a hijacked channel fast. Use YouTube's account-recovery flow for a hacked or hijacked channel through the official Help Center at support.google.com, and file the fraud with the FBI Internet Crime Complaint Center at ic3.gov and the FTC at reportfraud.ftc.gov. Warn your audience through any channel you still control so they do not fall for the crypto-giveaway stream.
Scan the sponsorship link before you download anything
SafeBrowz runs a 3-layer detection architecture: Local + APIs + AI. For this scam it works at three points in the chain, before you ever open the file.
- Layer 1 - Local detection: 60+ URL patterns and a 550+ brand database run inside the extension before the page renders. A domain that fuses a real brand name with sponsorship, offers, media-kit, or deal wording on a host the brand does not own matches instantly, so the fake partnership landing page is flagged on sight.
- Layer 2 - API checks: the media-kit download host and known malware-distribution domains are cross-referenced against Google Safe Browsing, PhishTank, URLhaus, and other threat feeds, plus a scam-TLD list. Stealer-delivery domains are frequently reported within hours of a campaign going live.
- Layer 3 - AI deep scan (Premium): content analysis in 100+ languages reads the page and flags fake YouTube or Google sign-in forms, and drainer-style "giveaway" content, on a host that is not accounts.google.com or youtube.com. Its limit is the one this article documents: it can only judge what the page actually contains, and the sponsorship platform described above carries a Sign in with Google button and a Google logo but no password field and no login form in its own markup, because the sign-in window is drawn only after you click.
Detection signatures are derived from threat-intelligence research and our internal brand database, not from user browsing data. SafeBrowz does not store per-user browsing history.
Honest scope: we did not catch the sponsorship platforms in this article on our own, and you should know that before you trust any of it. All four were returning clean on 14 October 2026, and they are on our block list because we read the research, verified the domains ourselves, and added them that day. We should be precise about which of those is actually damning: three of the four had already stopped serving the platform, so there was nothing on them left to judge. The one real miss is the live one. Two reasons, both worth understanding: the hostnames carry invented words like tubive, matchy and scouty rather than a brand name, so there is no brand string to match on sight, and the landing page holds no password field at all, because the counterfeit sign-in window is drawn by script only after you click. A lookalike domain and a visible fake login are things we are good at. An invented brand that hides its login behind a button is not, and no amount of our own marketing changes that. We also cannot un-send an email already in your inbox, and we cannot recover a channel that has already been hijacked: if the takeover happened, the recovery steps above and a fast report are your path back, not a scanner.
Test a sponsorship link before you download anything
Got a brand-deal email with a "media kit" or "brief" link? Paste the landing page or download link here first. Our 3-layer engine (Local + APIs + AI) returns a verdict in about 3 seconds. Free, no signup.
Frequently asked questions
Is this YouTube sponsorship email real?
If it is unsolicited and it requires you to download a media kit, brief, or contract before you can see the deal, treat it as a scam. Real brands discuss terms in the email body or a browser-viewable document and reach out from their verified corporate domain, such as nvidia.com or adobe.com. A required download, especially a password-protected .rar, .zip, or .exe, is the trap that installs an info-stealer. Verify any offer by contacting the brand through its official website, not by replying to the email or clicking its links.
Can they hijack my channel if I have 2FA turned on?
Yes, and this is the dangerous part. Two-factor authentication protects the login moment, but the info-stealer copies your active browser session cookie, the token that proves you already logged in and passed 2FA. The attacker loads that cookie into their own browser and lands inside your session with no password prompt and no 2FA challenge. The fix is not only 2FA: avoid the download entirely, and if you suspect infection, change your password and sign out of all sessions immediately, which invalidates the stolen cookie.
What does a real brand sponsorship email look like?
A genuine offer comes from the brand's verified corporate domain, references specifics about your channel rather than generic flattery, and does not force you to download an executable or archive to proceed. Terms, rates, and briefs are shared in the message body, a viewable document, or a call. There is usually a conversation before any paperwork, and a genuine brand never needs you to log in to your Google or YouTube account to prove the channel is yours. If any of those are missing, verify directly with the brand before doing anything.
I downloaded the brief. What should I do now?
Assume the stealer ran. From a different, trusted device, go to accounts.google.com, change your Google password to a long unique one, and then sign out of all other sessions and devices, which invalidates any stolen session cookie. Review your recovery email and phone, check channel managers and connected apps in studio.youtube.com, and run a full malware scan on the infected machine, rebuilding it if the channel is important. Then report the incident. Speed matters because a stolen cookie can be used within minutes.
The sponsorship email had no attachment. Does that mean it is safe?
No. A second version of this campaign, documented by ESET in October 2026 but running since at least June 2026, sends no file at all. It links to a polished sponsorship platform that greets you with your real channel name, subscriber count and video count, lines up a wall of brand logos, and asks you to confirm channel ownership by signing in with Google. It never sends you to Google. It draws a counterfeit browser window inside the page, with its own address bar reading accounts.google.com and a padlock icon, and puts the sign-in step in that frame. Your two-factor code is genuine and correct, and it is captured next to your password and used to sign in as you. So an attachment-free offer is not a safe offer, and an address bar you can see inside the page is not evidence of anything. Open a new tab and type the address yourself rather than signing in from any link.
How do I report a hijacked YouTube channel?
Use YouTube's official account-recovery flow for a hacked or hijacked channel through the Google Help Center at support.google.com, which can help restore access and revert malicious changes. File the fraud with the FBI Internet Crime Complaint Center at ic3.gov and the FTC at reportfraud.ftc.gov. If your channel was rebranded to run a crypto-giveaway stream, report that stream through YouTube as well, and warn your audience through any account you still control so they do not send money to the scam.
Do real brands ask you to download a .rar or enable macros?
No. A legitimate brand partnership never requires you to open a password-protected archive, run an .exe disguised as a document, or enable macros to view a brief. Those steps exist only to bypass antivirus and execute malware. If a sponsorship email leans on any of them, it is a phishing lure carrying an info-stealer. Delete it, and if you want to be sure the brand is not really trying to reach you, contact the company through its official website.
Bottom line: This scam works because it hides malware inside the everyday paperwork of a brand deal. It borrows a real brand's name, sometimes a deepfaked executive, and the ordinary excitement of a sponsorship offer, then asks for the one thing that unravels everything: that you open a file. Once the stealer has your session cookie, your 2FA never gets a chance to fire, and the channel you built becomes a stage for a crypto scam aimed at your own fans. The defense is steady and does not change with the brand. Never download a brief or sign in from an unsolicited sponsorship email, verify the offer on the brand's real site, and add a browser-layer scanner like SafeBrowz so the fake page is flagged before it can ever hand you the file.