A password attack is an attempt to get into an account with a stolen, leaked or guessed password, rather than by tricking the owner into typing it on a fake page. The common forms are credential stuffing (passwords leaked in other breaches), password spraying (a few common passwords tried across many accounts) and brute force (systematic guessing).
Microsoft’s Digital Defense Report 2025 says identity-based attacks rose 32% in the first half of 2025, and more than 97% of them were large-scale password attacks (Microsoft). The password attack examples below hit a ride-hailing company, a payments firm, a DNA testing site, cloud data customers and Microsoft itself.
How a password attack works
The attacker starts with a list of passwords, which comes from one of three places:
- Breach dumps and dark-web lists, for credential stuffing (OWASP).
- Lists of the most popular passwords, for password spraying, tried slowly through other people’s home internet connections (residential proxies) so no account locks out (Microsoft).
- Infostealer malware logs from infected personal or contractor laptops, sold on criminal markets (Mandiant).
Scripts try them at website and mobile-app logins, Microsoft 365, Citrix and VPN portals (CISA), and SaaS tools, on Windows, macOS, Linux, network devices and cloud services alike (MITRE). Without MFA, a correct password is all an attacker needs.
When the account has push MFA, attackers add MFA fatigue (also called MFA bombing or push bombing): they trigger approval prompts again and again until someone taps “Approve”. The same flood works against Microsoft Authenticator, Duo and Okta Verify (MITRE), and Apple users have been bombarded with password-reset prompts on an iPhone, Mac and Apple Watch at once (Krebs on Security).
Why it works on people
People reuse passwords, so one old breach opens accounts somewhere else. And a stream of prompts wears people down: CISA says users approve “accidentally or from MFA fatigue”, or “assume one of the requests is legitimate” (CISA).
Tell-tale signs
- An MFA prompt, code or password-reset request the account owner did not start, especially several in a row.
- “New sign-in” emails or an unexpected lockout.
- A call or message from “IT” or “support” right after a burst of prompts.
- For admins: failed logins across many accounts, logins from proxies or Tor, and MFA devices registered from unfamiliar places (CISA).
Password attack examples
Uber MFA fatigue (2022, ride-hailing).
Uber says an attacker likely bought a contractor’s corporate password on the dark web after malware infected the contractor’s personal device. It counts as a password attack because Uber says the attacker started from a password bought on the dark web, not one typed into a fake page. The attacker “repeatedly tried to log in”, each attempt sent an MFA approval request, and eventually the contractor accepted one. In press reports the attacker claimed to have messaged the contractor on WhatsApp posing as Uber IT, a fake support contact that Uber’s statement does not mention. The attacker reached Slack and G-Suite and posted in a company-wide Slack channel, though Uber found no access to production systems or sensitive user data (Uber). A run of prompts nobody started means someone already has the password.
23andMe credential stuffing (2023, consumer genetics).
The attacker logged into about 14,000 accounts with passwords reused from other breaches. Through a relative-matching feature, those logins exposed about 5.5 million DNA Relatives profiles and about 1.4 million Family Tree profiles, belonging to people whose own passwords were never used (23andMe). The UK ICO fined 23andMe £2.31 million in 2025, citing no mandatory MFA and weak password rules (ICO).
Microsoft password spray (disclosed January 2024, software and cloud).
A Russian state-sponsored group sprayed passwords and got into a legacy, non-production test account that had no MFA. It then used an old test app’s elevated access to read Microsoft corporate email accounts (Microsoft). A forgotten test account needs MFA as much as a live one.
Snowflake stolen-password logins (2024, cloud data, many sectors).
Attackers logged into Snowflake customer accounts with old stolen passwords, without breaching Snowflake’s own systems or phishing anyone. Mandiant traced the passwords mainly to infostealer infections, some on contractor laptops also used for gaming and pirated software, and the stolen data was used for extortion. None of the accounts had MFA, and some passwords had not been changed for up to four years (Mandiant). An old stolen password keeps working until it is changed or MFA is put in front of it.
PayPal credential stuffing (2022, payments).
Attackers logged into customer accounts with stolen usernames and passwords. Tax documents in those accounts showed unmasked Social Security numbers, and PayPal notified 34,942 people (The Record). New York’s financial regulator fined PayPal $2 million because it “did not require customers to use multifactor authentication or use controls such as CAPTCHA or rate limiting” (NYDFS).
DraftKings credential stuffing (2022, sports betting).
Three men used credentials bought on the dark web to get into about 60,000 accounts. In about 1,600 they added their own payment method and withdrew the funds, about $600,000 in total (CyberScoop). All three went to prison (US Department of Justice). A stored payment method makes a low-value login worth stealing.
Roku credential stuffing (2024, streaming).
Two waves hit more than 15,000 accounts in March and more than 576,000 in April. In fewer than 400 cases, attackers bought subscriptions and hardware with stored payment methods. Roku reset passwords and turned on MFA for all accounts (CBS News). MFA only protects accounts that already have it when the attack starts.
How to prevent password attacks
Personal habits to prevent password attacks
- Use a separate password for every account, kept in a password manager, so a breach somewhere else opens nothing.
- Turn on MFA for every account that offers it, starting with email and any old or “test” account, where a guessed password does the most damage.
- Deny an MFA prompt or reset code nobody requested and report it, however many arrive (CISA).
- Change a password the same day a breach notice names it, along with every account that shares it.
- Sign in to work accounts only on work devices, because infostealer malware on personal and shared devices collects logins.
Security controls to prevent password attacks
- Enforce MFA on email, VPN, remote access and cloud admin accounts, with no forgotten old or test accounts.
- Move accounts to passkeys or security keys where possible, since they leave no prompt to approve and no password to reuse. Where push MFA stays, switch on number matching (Microsoft Learn).
- Put rate limiting, CAPTCHA and MFA on customer login pages, so large-scale password guessing hits a wall.
