Stealer Logs: What They Hold and How to Find Your Accounts
Skip to content

Stealer Logs

A stealer log is the package of data that infostealer malware copies from one infected device: saved logins with the web address each belongs to, browser cookies, autofill entries and details of the machine.

On this page

Criminals collect logs in bulk, sort them and sell them, and buyers search them by email domain and by sign-in address. Each log describes one device at one moment, so every entry in it has to be read against what has changed since.

This guide covers what a log contains, how logs are produced and traded, how an organization finds its own accounts in them, and what it does when a match appears.

Anatomy of a Stealer Log

A log usually arrives as a compressed folder for one device, with text files and a subfolder for each browser and application the malware read. The Australian Signals Directorate’s ACSC lists what infostealers can take in its advisory The silent heist, from passwords and authentication cookies to chat logs, browsing history and VPN credentials.

For a defender, each part of the log answers a different question.

Part of the log What it records What it tells the defender
Credential entries The sign-in address, the username and the password, grouped by browser and profile. Which system each login opens, and whether that system belongs to the organization, a supplier or its customers.
Cookie files The domain, name, value and expiry of each cookie the browser held. Which signed-in sessions could be resumed elsewhere, and until when.
Autofill entries Names, phone numbers, postal addresses, other email addresses and sometimes card numbers typed into forms. The other identities the same person uses, including personal email addresses.
System summary Computer name, local username, operating system, IP address, location, installed software and the time of infection. Who the device belongs to, whether the organization manages it, and when the review period starts.
Application data Logins saved by VPN, file transfer, email and chat clients. Access that never passed through a browser, such as a remote access client.
Files and wallets Documents from common folders and cryptocurrency wallet data. Data exposure that sits apart from account access.

The sign-in address stored with each password is what makes a log searchable. A defender can see the exact page a password was typed into, which ties a credential to a system as well as to a person.

Dates in a Log

A log carries several dates, and each one sets a different clock for the response.

  • Infection time: The system summary records when the malware ran, and that moment is the earliest point from which stolen data could have been used.
  • Cookie expiry: Each cookie carries its own expiry date, which shows how long a stolen session could stay usable if nobody ends it.
  • Password age: The directory records when each password last changed, and a change after the infection time makes the stolen password stale.
  • First seen: The monitoring source records when it found the log, which is often days or weeks after the infection.
  • Reappearance: The same log can surface again in later collections, sometimes months on, carrying its original infection time.

The gap between infection time and first seen is the period nobody was watching. Every review of sign-in records starts from the infection time, whatever date the alert arrived.

Reappearance explains why one device can produce several alerts. A computer name and local username that match an earlier log usually mean the same data being passed around again, and a new infection time on the same machine means it was infected again.

Session Cookies in a Log

Cookies matter because they carry a session that already passed sign-in. The ACSC advisory notes that some authentication cookies keep a user signed in for days, and that a stolen cookie could bypass multi-factor authentication requirements.

A cookie file holds cookies for every site the browser visited, most of them useless to an attacker. The ones worth checking belong to the organization’s own domains and services.

  • Identity provider cookies: A session cookie for the single sign-on service can open every application connected to it.
  • Application cookies: Business applications often keep their own session after sign-in, so ending the identity session may leave the application session running until it expires.
  • Administration consoles: Cookies for cloud, firewall, backup or remote management consoles give control of infrastructure for as long as they last.
  • Webmail and file storage: Cookies for mail and document services give direct access to data.

Microsoft documents that revoking a user’s sign-in sessions in Microsoft Entra ID invalidates the refresh tokens issued to applications and the session cookies in the user’s browser, with a delay of a few minutes. The same documentation states that it does not revoke sessions for external users, who sign in through their home tenant. A guest account found in a log needs a separate decision, covered in the supplier section below.

Our session hijacking guide explains how stolen sessions are replayed and what ends each kind.

How Stealer Logs Are Produced

Logs are the product of a supply chain. The ACSC describes infostealers sold as malware as a service for a monthly fee, with a dashboard that builds the malware, organizes stolen data and counts infected systems.

The people who spread the malware are called traffers in the ACSC advisory. They reach victims through phishing, malicious search results, malicious advertising, pirated software, links in social media and video descriptions, and fake browser updates. Most campaigns are opportunistic, and some are aimed at particular industries when buyers ask for access to them.

The scale of a single family shows how many logs one product can generate. Microsoft reported that it identified more than 394,000 Windows computers infected with the Lumma stealer worldwide between March 16 and May 16, 2025. It described Lumma as malware sold through underground forums since at least 2022, with customers tracking stolen information through an online portal.

  • Collection: The malware runs on the device, reads the browser stores and application files, and packages what it finds.
  • Upload: The ACSC notes that logs go to command and control servers, and that many infostealers share a feed of logs through messaging apps such as Telegram and Discord.
  • Cleanup: Many infostealers are able to delete themselves after the upload, according to the same advisory.

Law enforcement and industry action against a family interrupts new infections. In May 2025 Microsoft, acting under a US court order and alongside the Department of Justice, Europol and Japan’s JC3, seized or blocked around 2,300 domains used by Lumma. Logs collected before a disruption stay in circulation, so an organization’s exposure outlives the malware that caused it.

How Stealer Logs Are Traded

The ACSC states that specialized marketplaces exist on Telegram and across the dark web for the sale and trade of logs. A joint advisory from CISA and the FBI on LummaC2 cites private sector figures of more than 21,000 market listings selling its logs on criminal forums between April and June 2024.

Logs change hands in several forms, and the form decides how quickly an organization’s data moves.

  • Single logs: Markets offer logs one device at a time, so a buyer can choose devices that hold corporate logins.
  • Subscription channels: Channels publish batches of fresh logs to paying members, which spreads one device’s data to many buyers at once.
  • Archives: Older logs are bundled into large collections and passed around again, which is why old infections keep producing new alerts.
  • Extracted credential lists: The address, username and password lines are pulled out of many logs and merged into lists that feed automated sign-in attempts.

The ACSC describes the buyers. Initial access brokers purchase and validate stolen credentials, then auction them to criminals who use them to get into the organization’s corporate network. Other buyers exploit the victim directly, through identity theft and extortion.

For a defender, each form has a different meaning. A single log with a corporate login is a targeted risk to that account. A credential list puts the same password in front of every service that accepts it, which turns one person’s exposure into a credential stuffing risk wherever that password was reused.

Where Your Accounts Appear

An organization’s data appears in logs from devices it owns, devices its staff own, and devices belonging to people it has never met. Matching by email domain alone finds only part of it.

Exposure How it is matched Who owns the response Usual action
Staff on corporate sign-in pages A work address on the organization’s single sign-on, VPN or webmail address. Internal IT and security. Reset, revoke sessions and review sign-ins.
Staff on third-party services A work address on a supplier’s or software provider’s sign-in page. The organization, with the service owner. Reset on that service and check where the password was reused.
Suppliers and contractors Another company’s address on the organization’s own sign-in pages. The supplier, with the internal access owner. Disable or reset the account and notify the supplier.
Customers Any address on the organization’s customer portal or app sign-in page. Product, fraud or customer operations. Force a reset, end sessions and tell the customer.
Personal addresses on work systems A personal email address used to sign in to work storage or applications. The user, with internal IT. Handle as a staff exposure.
Shared and service logins Generic usernames such as admin or a service name on administration pages. The system owner. Rotate the credential and move to named accounts.

The last column depends on who controls the account. A staff account can be reset by the organization the same day, while a customer or supplier account needs someone else to act as well.

Judging Each Match

A first search often returns entries that are years old, belong to departed staff or point to services the organization no longer uses. Each match is judged on a few questions before anyone acts.

  • Currency: The stolen password is compared with the account’s last password change, and a confirmed leaked credentials detection settles the question for directory accounts.
  • Account state: The account is checked against the directory and HR records to see whether it is active, disabled or belongs to someone who has left.
  • Reach: The sign-in address shows what the login opens, and a single sign-on or administration login outranks an individual application.
  • Live sessions: Cookies for the organization’s domains with expiry dates still ahead make the match urgent even when the password has changed.
  • Device ownership: The computer name in the system summary shows whether the device is managed, which decides what endpoint response is possible.
  • Repetition: A device or user seen in earlier logs is linked to the earlier case, so a resale is recorded as such and a new infection opens a new case.

The answers place each match in one of a small set of states.

  • Current and privileged: The password works or a session is live on an administrative or single sign-on account, so the response starts within the hour.
  • Current: The password or session is valid for an ordinary account, so the response starts the same day.
  • Stale: The password changed after the infection and no session can still be live, so the match is closed after a check for password reuse.
  • Out of reach: The account belongs to a supplier, a customer or a service the organization does not run, so the match goes to its owner with the evidence.

Customer Accounts in Logs

A search by sign-in address finds the organization’s customers as well as its staff. Their devices are out of reach, yet the accounts are on the organization’s own service and the organization holds the controls.

Customer matches usually arrive in volume. A popular portal can appear in thousands of logs, because each infected household computer that ever signed in to it adds an entry.

  • Force a reset: The account is marked so that the next sign-in requires a new password, chosen on a clean device.
  • End sessions: Existing sessions and remembered devices for the account are cleared.
  • Step up sensitive actions: Changes to payment details, delivery addresses or contact details require a second factor or a fresh sign-in for a period.
  • Tell the customer: The message explains that the customer’s own computer was likely infected, without asking them to click a link, and points to how to sign in safely.
  • Watch for automated sign-ins: Matched accounts are watched for attempts from hosting providers and for bursts of failed sign-ins, the signature of automated credential testing.

Standards support this approach. NIST SP 800-63B-4, the Digital Identity Guidelines for authentication, says verifiers shall compare new passwords against a blocklist that includes compromised passwords, and shall force a change when there is evidence a password has been compromised.

Suppliers, Contractors and MSPs

A log from a supplier’s device can hold logins to the organization’s systems, and the organization learns of it only if it searches by its own sign-in addresses. The ACSC advises organizations to review the supply chain risk from vendors that access their networks, including software-as-a-service providers and managed service providers.

Managed service providers carry particular weight. A technician’s log may hold logins for remote management tools, client portals and administration consoles across many client organizations at once.

  • Contain on your side: Disable or reset the supplier’s account on the organization’s systems, and end its sessions where the organization’s own controls allow.
  • Guest accounts: Microsoft’s session revocation does not reach external users who sign in through their home tenant, so the guest account is blocked locally and the supplier is asked to reset at home.
  • Notify the supplier: The supplier’s security contact receives the account, the sign-in address, the infection time and the device details from the log.
  • Ask for confirmation: The supplier confirms when the device has been cleaned and the account reset, before access is restored.
  • Review the contract: Terms for remote access can require managed devices, multi-factor authentication and prompt notice of infections.

One Exposure Followed Through

One match at acme-corp shows how a log entry becomes a closed case.

  1. The match. Monitoring for acmecorp.com and acme.com.au returns a log containing an accounts payable officer’s work address on the single sign-on page, a VPN portal password, a payroll application login, and an administration login for the acme.com.au customer portal.
  2. The system summary. The computer name and installed games point to a personal laptop, and the infection time is nine days before the alert. The device is out of the IT team’s reach, so the response runs entirely through accounts.
  3. Currency. The directory shows the password last changed five months earlier, and Microsoft Entra ID has raised a leaked credentials detection for the same user. The cookie file holds a single sign-on cookie that expires in twelve days.
  4. Containment. The security team blocks sign-in, resets the password from a managed device and revokes sessions. The payroll provider resets the payroll login, and the portal team rotates the administration password and moves it into a named account.
  5. Use. The non-interactive sign-in records show token activity from a hosting provider two days after the infection. The case is escalated to an account takeover investigation, which finds a new inbox rule and removes it.
  6. Closure. The officer stops using the laptop for work until it is reinstalled. Three weeks later the same log reappears in an archive with the original infection time, and it is linked to the closed case.

The case started from one alert and reached systems with different owners. The single sign-on cookie, which the directory alert could not have shown, was the part already in use.

Evidence for Each Match

Each case keeps a record of what was found, what was checked and where the answer came from. The table lists the records that settle most cases in a Microsoft 365 environment.

Record Question it answers Where it is kept
Log entry The sign-in address, username, infection time and device details. The monitoring source’s finding, saved with the date it was first seen.
Leaked credentials detection Microsoft confirmed the stolen password matches the current one. Entra ID Protection risky users report, or risk detections in Microsoft Graph with the type leakedCredentials.
Password last changed Whether the stolen password predates the latest change. The lastPasswordChangeDateTime property of the user in Microsoft Graph.
Interactive sign-ins Password use from new addresses, devices or countries. Entra ID sign-in logs, interactive user sign-ins.
Non-interactive sign-ins Token and session use without a fresh sign-in. Entra ID sign-in logs, non-interactive user sign-ins.
Session revocation The time existing sessions were cut off. The signInSessionsValidFromDateTime property of the user, and the directory audit log.
Remote access records VPN or remote desktop connections with the account. The VPN gateway or remote access service logs.
Third-party sign-ins Use of the other logins in the log. Each application’s administration audit log.
Device ownership Whether the computer in the log is managed. The endpoint management inventory, searched by computer name.
Case record Who did what and when. The ticket or case management system.

Other identity providers keep the same records under other names, so each claim in the case has a log behind it.

Acting on a Match

The response to a match depends on its owner, and the owner table above decides who acts. For staff whose devices were infected, the device steps and the reset and revocation order are set out in our infostealer response guide.

Around those steps, a log match needs some handling of its own.

  • Work the whole entry: Every address in the log is reviewed, including third-party and personal services, because a password reused there may also work on a work system.
  • Fix every owner’s part: Matches for suppliers, customers and third-party services are handed to their owners with the evidence on the same day.
  • Retire shared logins: A shared or generic credential found in a log is rotated and replaced with named accounts, because its exposure cannot be traced to one person.
  • Start reviews at the infection time: Sign-in and audit reviews cover the period from the infection time, which is often well before the alert.
  • Watch for the same device: The computer name, local username and account are kept on a watch list, so a new infection is told apart from a resale.
  • Tell the user what was found: The user hears which device, which logins and what to do on personal accounts, in plain words and without blame.

A case closes when every current credential has been reset or rotated, every live session has ended, and the review of use has an answer. Where that review finds use, the case becomes an intrusion and is handled as one.

Standing Up Log Monitoring

A stealer log program can start with existing staff and one monitoring source. The first six weeks or so are mostly inventory and agreement, and they fall into stages.

  1. The keys. The security lead lists every email domain, sign-in address and customer-facing sign-in page, and names an owner for each.
  2. The backlog. The first search returns historical matches, and the team sorts them by currency and reach, closing stale ones with a recorded reason.
  3. The playbooks. Each owner type gets a written response: staff, suppliers, customers and shared logins, with a supplier notice and a customer message ready to send.
  4. Steady running. New matches are judged within a day, and the case records feed a short monthly report.

A few measures show whether the program works.

  • Time to reset: The days from first seen to a completed reset, reported separately for staff, suppliers and customers.
  • Share of current matches: The proportion of matches with a valid password or live session, which shows how fresh the sources are.
  • Repeat devices: The count of devices seen again with a new infection time, which points to staff who need help with a personal computer.
  • Use found: The number of cases where the review found the stolen access had been used.

The same period is a good time to tighten controls that limit what a log is worth. The ACSC recommends that work credentials stay out of personal password managers, including the browser’s own, unless the employer has approved it, and that session tokens and cookies have enforced lifetimes.

Stealer Log Monitoring With ThreatDefence

ThreatDefence reads stealer logs as they surface and matches each record to a customer’s domains, its staff accounts and the customer accounts on its services. These records jump the queue of leaked logins, since a cookie taken from a signed-in browser lets someone resume the session with no password and no second factor.

For environments ThreatDefence monitors, the 24x7 SOC takes a staff match through to a closed account. The username is tied to a directory account, the password is tested where the customer’s integrations permit, and analysts go through sign-in records from the date of the leak, looking for places, devices and networks the user has never used. The reset and the session revocation then follow the customer’s procedures, and evidence of use opens an investigation.

The computer named in the log tells the SOC whether the customer manages that machine, and the response differs accordingly. When an old log is resold in a new collection, it is attached to the earlier case and no fresh alert is raised. The customer gets a single record per exposure, and the brand monitoring page describes the service around it.

Frequently Asked Questions

Which files make up a stealer log?

A stealer log holds what the malware found on one device: saved logins with the sign-in address for each, browser cookies with their expiry dates, autofill entries, and a system summary with the computer name, IP address and infection time. Many also include logins saved by VPN, file transfer and chat clients, documents from common folders and cryptocurrency wallet data.

How are stealer logs sold?

Logs are sold individually on criminal markets, published in batches to paid subscription channels, and bundled into archives that circulate for months. The ACSC notes that specialized marketplaces for logs exist on Telegram and across the dark web. Initial access brokers buy logs, test the corporate logins in them and auction the working ones to criminals who use them to enter the organization’s network.

Which search terms find our accounts in stealer logs?

Search by every email domain the organization uses and by the full host names of its sign-in pages, VPN portals, administration consoles and customer portal. Exposure monitoring services run this search continuously. In Microsoft 365, Entra ID Protection’s leaked credentials detection adds a confirmation that a directory password found in leaked data is still current.

Does Microsoft Entra ID warn about leaked passwords?

Yes. Entra ID Protection raises a leaked credentials detection when it finds a user’s valid credentials in leaked data and confirms them against the tenant’s current password hashes. Microsoft lists it as available without Entra ID P2. On-premises passwords are covered only with password hash synchronization, and the detection does not report stolen session cookies or logins for other services.

Is an old stealer log still a risk?

Yes, in some cases. A log stays in circulation long after the infection, so its age alone settles nothing. If the password has changed since the infection time and no session cookie can still be live, the main remaining risk is the same password reused on other services. If the password never changed, the log is as dangerous as on the day it was taken.

How long does a stolen session cookie stay useful?

A stolen cookie works until it expires or the service ends the session. The expiry is recorded in the log, and the ACSC notes that some authentication cookies keep a user signed in for days. Revoking sessions in the identity provider ends single sign-on sessions, while applications that keep their own sessions may need them ended separately.

What should we do when a customer's login appears in a stealer log?

Force a password reset at the next sign-in, end existing sessions and remembered devices, and require a second factor for sensitive changes for a period. Then tell the customer their own computer was likely infected, without including a link to click. NIST SP 800-63B-4 says verifiers shall force a password change when there is evidence of compromise.

Should a supplier be told when its staff appear in a log?

Yes. The supplier’s device holds the infection, and only the supplier can clean it and reset the account in its own systems. Send its agreed security contact the account, the sign-in address, the infection time and the device details, and keep the account disabled on your side until the supplier confirms both steps are done.

Can stealer logs be removed from criminal markets?

No. A log is copied to buyers, channels and archives as soon as it is sold or published, and no single takedown reaches every copy. The response is to make the data worthless: reset or rotate every current credential, end every live session, and keep watching for the same device and account in later collections.

Talk to Our Experts

See how ThreatDefence automates security monitoring, compliance and cyber risk management in one platform.

    I'm an MSP, MSSP, or reseller

    By submitting this form, you accept the terms in our privacy policy.