SeverVault: Infostealer Logs Sold as AI Victim Profiles
On November 14, 2026, a seller on DarkForums opened a thread for SeverVault, a shop offering infostealer logs and combo lists. The pitch leans on scale and automation: an inventory said to run into the tens of millions of logs, more than a billion parsed accounts, entry prices of $0.25 and free registration. The headline feature, though, is that each log arrives with an AI-written summary of the machine it was stolen from.
That summary changes what a buyer is paying for. A stealer log has always been a pile of saved passwords, cookies, autofill entries and system details that someone had to sift through by hand. SeverVault does the sifting in advance and presents the result as a profile: who appears to use the computer, where they seem to work and which of their accounts might open a door into an organization.
For security teams, the takeaway is uncomfortable. An infection on a government worker’s home computer or a bank employee’s family PC no longer sits anonymously in a bulk dump. It arrives labeled with the employer, the sector and the kind of corporate access it may hold, from government portals to Citrix sessions.
Key Takeaways
- SeverVault is an infostealer log and combo list shop promoted on DarkForums, with an advertisement dated November 14, 2026
- Every log reportedly comes with an AI-generated description of the infected machine and its likely owner, plus tags naming organization domains, SSO providers, banks and crypto wallets
- Listings seen so far tie $9 to $10 logs to the home computers of government employees and to households with corporate Citrix and SSO access, including a defense contractor and a bank employee
- Inventory size, victim profiles and claims about working tokens all come from the seller and have not been verified
- Exposed logs should be handled as endpoint compromises, with session and token revocation alongside password resets
SeverVault at a Glance
| Attribute | Details |
|---|---|
| Name | SeverVault |
| Type | Infostealer log and combo list shop |
| Promoted On | DarkForums |
| Advertisement Date | November 14, 2026 |
| Claimed Inventory | Tens of millions of logs, 1,000,000,000+ parsed accounts |
| Entry Price | $0.25 (claimed) |
| Differentiator | AI-written profile, tags and score per log |
| Operators | Unknown |
| Stealer Families | Vidar and Rhadamanthys seen in listings |
From Bulk Dumps to Searchable Victims
The SeverVault advertisement promises that AI has gone through every log to pull out card BINs, shopping accounts, crypto wallets, email addresses and passwords, then written a description for each one. Registration is open to anyone, and the operators publish four clearnet domains and four onion mirrors to keep the shop reachable.

Shops that index stolen data by location or account type are nothing new. What SeverVault adds is a layer of interpretation on top of the raw files. Instead of a country code and a browser count, a buyer sees a headline such as “The Defense Contractor and Student Household Dump” and a paragraph explaining why the machine is worth buying.
Inside the Storefront

Each listing in the storefront pairs the AI profile with the metadata stealer log buyers already expect:
- Stealer family: the malware that produced the log, with Vidar and Rhadamanthys among the listings seen
- Host details: operating system and build, country, state, city and ISP
- Profile: a headline, a short AI-written narrative, organization and service tags and a score out of 100
- Inventory counts: the number of links, saved passwords and cookies in the log
- Sale details: date, file size, a masked seller ID with a tier badge and a price
The tags read like a target summary: organization domains, labels such as “government official” or “student,” SSO providers such as Okta, Auth0 and GitHub, crypto wallets such as Exodus, MetaMask and Trust Wallet, PayPal and other bank accounts, and gaming platforms. The shop gives no explanation of how scores are calculated; the listings reviewed for this post ranged from 54 to 100.

Which Victims Stand Out
The listings that stand out are the ones whose profiles point to a government employer or to corporate Citrix and SSO access, the machines that become most valuable once they carry a profile.
Government and Defense-Linked Machines
Several listings are framed around government employees. A Los Angeles County official’s workstation reportedly holds active sessions on government portals such as SSA, USCIS and EDD, along with a private RDP key. Others include a Virginia government official described as an active job seeker and a DC government employee’s family computer with session cookies for DC government portals. They sit next to “The Defense Contractor and Student Household Dump,” a Maryland machine whose primary user is described as a defense contractor with corporate SSO and Citrix access. Most are described as household or family machines rather than work devices. The profile is what connects a personal computer to an organization worth targeting.

Financial Sector Households
One listing, titled “The Bank Employee Family’s Shared Workstation,” describes a Windows 11 machine in a Texas suburb used by a financial sector employee, a minor student and a sibling. According to the profile, it holds corporate SSO and Citrix session traces, saved credentials for major US banks, retirement portals and payroll services, live Steam and Discord tokens and eight Google OAuth refresh tokens. The Vidar log was listed for $9 with a score of 100/100.
Why Profiled Logs Raise the Stakes
Personal and shared machines have long been a blind spot. They sit outside corporate endpoint controls, yet they often store work email sessions, VPN or Citrix logins and SaaS cookies. In a bulk dump, that exposure is easy to miss among millions of consumer logs. With an AI profile attached, the same log is labeled with the employer and the access it holds, and priced accordingly.
The seller’s profiles may be wrong, inflated or built from a mix of users on one machine, and nothing published so far measures their accuracy. Still, a profile does not need to be perfect to save an attacker time. Narrowing tens of millions of logs to a short list of bank employees, defense contractors or government staff is enough to turn opportunistic theft into targeted initial access.
What a SeverVault Log Can Contain
Based on the listing profiles, a log may include:
- Saved credentials for banks, retirement portals, payroll services and investment platforms
- Session cookies for government portals, job boards, retail and streaming services
- Corporate SSO traces, with tags for Okta, Auth0, GitHub, Jira and Workday
- Citrix session traces
- Discord, Steam and Google OAuth refresh tokens
- Crypto wallet data, tagged for Exodus, MetaMask and Trust Wallet
- Payment card entries, autofill data and, in one case, a private RDP key
Session cookies and OAuth refresh tokens are the most immediate risk. A stolen, still-valid session can be replayed without passing through an MFA prompt, so revoking sessions belongs in containment alongside password resets.
SeverVault Infrastructure
Addresses published in the DarkForums advertisement, defanged:
Clearnet Domains
- sever[.]to
- sever[.]ws
- sever[.]cx
- sever[.]st
Onion Addresses
- severgoodb3kdzmdnrnx3gb4jbw4gxe43mhvnxaacfq7oowb2ttsh7id[.]onion
- sever4apt73wclanpzgoelsscbj274xupr4npybhyr4vyvxl7jlarrqd[.]onion
- severown5xd6ojimw7qij32dvgeav5ath7fei4w7rbui5zo2rutk3yad[.]onion
- severxpppwio5sku5ffx2ddlbqszqous3wr7hph52dlcbdr3henfibid[.]onion
How to Respond to Stealer Log Exposure
Reduce the Value of a Stolen Log
- Limit remote access to managed devices: require device compliance checks for Citrix, VPN and SaaS sign-ins so a family laptop cannot carry working corporate sessions
- Review shared devices: treat family computers and other shared machines used for work access as high-exposure endpoints with the same monitoring as staff laptops
- Shorten session lifetimes: reduce how long a stolen cookie or token remains usable
When a Log Surfaces
- Map the exposure: list every account, application and token the log touches, including remote access, SaaS and API platforms
- Clean the machine first: remove the infection or rebuild the device before issuing new credentials
- Reset and revoke together: change stored passwords and invalidate active sessions, refresh tokens and API keys in the same step
- Hunt for use: check authentication and audit logs for sign-ins, session reuse or API activity since the infection date
- Keep watching: monitor Dark Web shops and stealer log channels for your domains and employee accounts so new exposures are caught early
Frequently Asked Questions
What Makes SeverVault Different From Other Stealer Log Shops?
SeverVault attaches an AI-written profile to every log describing the infected machine, its likely owner and the access it holds, along with tags and a score. A buyer can pick a log tied to a specific employer, sector or type of access without opening the files first.
Is SeverVault’s Inventory Really Over a Billion Accounts?
That figure comes from the seller’s DarkForums advertisement, along with the claim of tens of millions of logs. Neither number has been independently confirmed, and seller claims on Dark Web forums cannot be taken at face value.
Can a Stolen Session Cookie Bypass MFA?
Yes, if the session is still valid. A cookie captured after the user has already signed in can be replayed by an attacker without a new MFA challenge, so password resets alone do not close the gap. Active sessions and tokens need to be revoked as well.
Why Would Attackers Want Logs From Home Computers?
Employees often sign in to work email, VPN, Citrix or SaaS tools from personal devices that sit outside corporate security controls. When a profile labels such a machine with its owner’s employer, an ordinary home infection becomes a possible entry point into that organization.
What Should We Do if an Employee’s Machine Appears in a SeverVault Listing?
Identify every account and application the log exposes, clean or rebuild the device, then reset passwords and revoke sessions and tokens at the same time. Afterward, review authentication and audit logs for any activity that suggests the access has already been used.

