Phishing-as-a-service (PhaaS) has spent years borrowing the manners of legitimate software. There are subscriptions, dashboards, product updates, customer support channels, and tidy automation. The key difference is that the “customer success” story ends with someone else losing control of a Microsoft 365 account.
BigBear 2.0 shows how mature that model has become. In September 2026, CloudSEK published findings from access to an active, multi-tenant phishing panel. The operation combined adversary-in-the-middle infrastructure, credential and session-cookie theft, affiliate access, Telegram automation, residential proxies, and a management interface that kept the machinery running at scale.
This isn’t a story about one especially clever phishing page. It’s a story about a service business built to produce them continuously, and that distinction changes what defenders should measure.
If a panel can replace domains, rotate infrastructure, and serve several operators at once, counting detections is bookkeeping. The metric that changes victim exposure is how quickly the campaign gets disrupted.
BigBear 2.0 by the numbers
CloudSEK’s investigation (opens in new tab) traced BigBear 2.0 after researchers gained visibility into its administrative panel. The data offered an unusually direct look at a live PhaaS operation rather than a collection of phishing pages viewed from the outside.
At the time of the report, the panel showed:
- 5,137 credential records: The panel had collected thousands of victim entries, creating a reusable store of identities for follow-on abuse.
- 4,148 session cookies: These tokens can let an attacker step into an authenticated session after the victim completes MFA.
- 1,032 plaintext passwords: Familiar credential theft remained part of the package, even as the operation expanded into session hijacking.
- 474 completed MFA-bypass authentications: The panel recorded hundreds of cases in which the workflow progressed beyond password capture.
- 461 targeted organizations: Victims spanned multiple sectors, with IT services and managed service providers making up the largest group.
- 3,331 unique victim IP addresses across more than 40 countries: The operation’s reach was international, and its infrastructure reflected that.
- At least five affiliate operators: Separate users accessed the service through Telegram bots and shared infrastructure, turning one platform into several active campaigns.
While BigBear 2.0 operates at a global scale, the telemetry (opens in new tab) reveals a heavily concentrated geographic and sector footprint:
- India emerged as the primary victim hotspot, accounting for 658 exfiltrated records (12.8% of the total dataset)
- Followed closely by France (463 records) and Saudi Arabia (353 records)
Beyond IT services and managed service providers (MSPs), the campaign’s high-density targeting extends directly into critical supply chain verticals, most notably pharmaceuticals, oil and gas, and consulting. This exposes high-value enterprise accounts to follow-on session hijacking and monetization by initial access brokers.
Those numbers are a snapshot, not a lifetime achievement award. The panel was active during the investigation, and the operators had already shown a willingness to delete and rotate infrastructure. A dashboard total tells defenders what the service has done. Its operating model explains what it can do next.
How a live PhaaS panel works
BigBear 2.0 packages the fiddly parts of phishing into a managed system. Affiliates don’t have to build a reverse proxy, engineer an MFA-bypass flow, manage servers worldwide, or hand-sort every captured session. They can rent access and concentrate on finding victims.
Affiliates rent the machinery
The panel supported administrator and user roles, which is a strong clue about the business structure. A central operator maintained the platform while affiliate users launched or managed their own activity. Telegram bots delivered data and coordinated access.
That arrangement lowers the barrier to entry. An affiliate can bring a lure, a target list, or a traffic source. The service provides much of the technical middle: phishing templates, infrastructure, token capture, telemetry, and a place to review the haul. Specialization makes the whole market more efficient, a sentence nobody enjoys writing about crime.
The reverse proxy sits in the login flow
BigBear 2.0 used Evilginx2-style adversary-in-the-middle infrastructure to target Microsoft 365. The victim saw a familiar login experience, but the phishing server relayed traffic between the browser and the real service.
The victim entered a username and password. Microsoft requested MFA. The victim completed it. Then the phishing server captured the authenticated session cookie returned after that successful check. MFA worked as designed; the attacker stole the proof that it had already been completed.
That’s why this attack belongs in the conversation about account takeover, not just credential harvesting. The prize is a live session with access to email, files, contacts, and whatever connected workflows the account can reach.
Automation turns stolen sessions into a queue
Captured data flowed through the panel and Telegram, reducing the delay between victim action and attacker replay. Custom JavaScript reportedly disabled or steered around certain protections, blocked some telemetry, and pushed the login flow toward options useful to the attacker.
Residential proxies spanning 69 countries (opens in new tab) helped affiliates make access appear geographically plausible. The platform integrates an automated, geo-matching residential proxy network that lets attacker sessions emerge near each victim, systematically bypassing Microsoft Entra ID’s “Impossible Travel” and location-anomaly detection rules. An attacker could follow a victim in one region with a session emerging nearby, making crude “impossible travel” logic less dependable.
Infrastructure is treated as replaceable inventory
CloudSEK identified 42 virtual private server nodes associated with the campaign. Twenty-six had been deleted since late July, suggesting rotation and counter-forensics rather than careful preservation of criminal antiques.
Domains and servers are disposable inputs in this model. Once an asset draws attention, the operator can retire it and point affiliates toward a fresh one. That’s why the surrounding relationships matter: shared certificates, hosting patterns, redirect behavior, scripts, templates, bot tokens, and other identifiers can connect today’s node to tomorrow’s replacement.
Convenience is the product for criminals
PhaaS economics are simple. Building an MFA-bypass operation from scratch takes skill, time, infrastructure, and maintenance. Renting one turns a capital project into an operating expense.
The platform operator benefits from recurring revenue and scale. Each new affiliate can generate more traffic without requiring a new core product. Affiliates benefit from faster launches and fewer technical headaches. When Microsoft changes a login flow or a browser behavior breaks the kit, one operator can update the service for everyone.
The package typically compresses several jobs into one subscription:
- Infrastructure management: Domains, proxies, servers, certificates, and routing are arranged or integrated for the affiliate.
- MFA-aware phishing flows: The kit relays legitimate authentication steps and captures the session afterward.
- Victim and campaign tracking: Dashboards organize credentials, cookies, IP addresses, targets, and conversion status.
- Exfiltration and alerts: Telegram delivery tells an affiliate when a useful session is ready, because apparently crime also suffers from notification fatigue.
- Evasion updates: Scripts, templates, and traffic controls can be revised centrally as defenses catch up.
This structure spreads capability beyond the small group that can engineer it. It also creates concentration. Affiliates may look separate at the lure level, yet rely on the same panel, scripts, infrastructure patterns, or control channels. That concentration gives defenders leverage if they can see it.
Why MFA completion isn’t the victory screen
MFA still stops enormous amounts of account abuse, and organizations should use it. BigBear 2.0 doesn’t make multi-factor authentication pointless. It shows why the factor and the surrounding authentication flow both matter.
Phishing-resistant methods such as passkeys and hardware-backed FIDO2 credentials are harder to relay because authentication is bound to the legitimate site. Yet BigBear’s reported use of scripts to suppress or steer users away from FIDO2 is revealing. The operator understood that stronger authentication was an obstacle and modified the experience to push users toward a weaker path.
Security teams should watch for successful authentication followed by suspicious session behavior, token reuse, new mailbox rules, unusual OAuth consent, unfamiliar device registration, and rapid access to sensitive resources. When a session is suspected, resetting the password alone may leave the stolen token alive. Revocation has to include active sessions and refresh tokens.
Detection volume is a terrible finish line
A dashboard can show 300 malicious URLs detected this month and still conceal an uncomfortable truth: the same panel kept operating throughout. Volume proves that sensors are seeing activity. It doesn’t prove the operation has lost reach.
For an always-on service, defenders need metrics tied to interruption:
- Time to validate: How long does it take to confirm that a reported domain, page, profile, or ad belongs to a malicious campaign?
- Time to contain internally: How quickly are URLs blocked, sessions revoked, affected accounts investigated, and users warned?
- Time to disrupt externally: How long does the domain, hosting node, fraudulent profile, ad, or other public asset remain available after confirmation?
- Campaign recurrence: When an asset disappears, how quickly does the operation return, and can you tie the replacement to known infrastructure before it collects another batch of victims?
- Downstream exposure: How many employees, customers, partners, or third parties could still reach the campaign while the response is underway?
These measures turn takedown into an operational discipline rather than a closing ticket. Faster validation and disruption reduce the window in which affiliates can monetize their rented access.
Campaign-level threat visibility also helps teams avoid whack-a-mole response. One domain may be the visible asset, while the panel, affiliate bot, proxy pool, and neighboring domains keep the campaign healthy. Map the cluster so defenders can prioritize the pressure point with the greatest effect.
What defenders can do against an always-on panel
No single control closes this market, but a coordinated program can make each victim more expensive and each campaign less durable.
- Use phishing-resistant authentication where possible. Prioritize passkeys or hardware-backed FIDO2 for administrators, finance, IT support, executives, and other high-impact roles.
- Revoke the full session after suspected phishing. Invalidate session and refresh tokens, review sign-ins, and inspect post-compromise activity instead of stopping at a password reset.
- Harden identity telemetry. Alert on suspicious cookie replay, new inbox rules, unexpected device or MFA changes, risky OAuth grants, and access patterns inconsistent with the user’s normal work.
- Find the impersonation infrastructure early. Monitor lookalike domains, paid ads, social accounts, messaging channels, and cloned login pages that funnel victims toward the panel.
- Connect assets into campaigns. Doppel’s Threat Graph links related infrastructure and attacker behavior, helping teams move from isolated indicators to the operation behind them.
- Disrupt beyond the first URL. Coordinate takedowns across domains, hosts, registrar relationships, fraudulent profiles, ads, and other reachable nodes, then watch for reappearance.
- Rehearse the whole response. Identity, security operations, legal, brand, communications, and service-desk teams should know who revokes access, who contacts providers, and who warns potential targets.
- Prioritize supply chain exposure. Pharmaceutical, energy, and consulting organizations should treat stolen sessions as potential initial-access inventory, since compromised accounts can be resold to follow-on actors.
The post-list work matters as much as the controls themselves. Review what survived, which assets reappeared, and how quickly the affiliate changed tactics. Every campaign is also a product-feedback session for the attacker; defenders should learn at least as quickly.
BigBear is a business. Make it a bad one
BigBear 2.0 is useful to criminals because it removes delay. It helps affiliates launch faster, receive stolen sessions sooner, and replace infrastructure with less effort. A strong defense attacks that advantage.
Doppel’s brand protection platform helps discover impersonation assets, connect them through shared campaign signals, and drive disruption before the operator gets comfortable. The objective isn’t to admire a growing pile of alerts. It’s to shorten the time the service can reach victims and increase the cost of taking it back down.
How long could a rented phishing panel keep targeting your brand after your team detected its first domain? Schedule a Doppel demo to see how graph-driven detection and takedowns can turn an always-on operation into a frequently interrupted one.