Back to Intelligence

Uber Third-Party Breach Exposes 150,000 Driver Records — Detection and Third-Party Risk Defense Guide

SA
Security Arsenal Team
November 16, 2026
10 min read

Uber has disclosed a data breach impacting approximately 150,000 individuals after attackers compromised a third-party services provider and stole names, physical addresses, and driver's license numbers. The critical detail here is not that Uber's own infrastructure fell — it is that the breach originated in the extended enterprise: a vendor that held sensitive personal data on Uber's behalf.

This is the recurring pattern of the modern breach economy. In 2025 and 2026, we've watched adversaries consistently route around hardened corporate perimeters by targeting the weakest link in the data supply chain — marketing platforms, claims processors, SaaS integrations, and managed service providers. The MOVEit cascade, the Ticketmaster/Santander Snowflake incidents, and now this Uber vendor compromise all share the same DNA: your data, someone else's security controls, your regulatory and reputational liability.

For defenders, the implications are immediate. Driver's license numbers combined with names and addresses are high-quality identity fraud material — sufficient for SIM-swap pretexting, synthetic identity construction, and targeted social engineering. If your organization shares PII with any external processor, this incident is your tabletop exercise made real.

Technical Analysis: Anatomy of a Third-Party Data Compromise

What was taken: Names, addresses, and driver's license numbers of roughly 150,000 people. No CVE is associated with this incident — this is not an unpatched-vulnerability story. It is an access-and-exfiltration story, which means the kill chain is behavioral and therefore detectable.

Attack chain (typical for vendor-side PII theft):

  1. Initial access at the vendor — credential phishing, exposed cloud storage, stolen session tokens, or a compromised SaaS account with access to the data store.
  2. Discovery — the attacker enumerates databases, object storage buckets, or file shares containing the contracted dataset. Expect bulk read patterns: full-table SELECT statements, S3 GetObject storms, or mass exports from a CRM/ticketing platform.
  3. Collection and staging — PII is aggregated into compressible archives. We consistently see 7z, rar, or zip utilities invoked against export directories, frequently with password protection to defeat DLP content inspection.
  4. Exfiltration — bulk transfer over HTTPS to attacker-controlled infrastructure, cloud storage (MEGA, Dropbox, paste sites), or via the vendor's own legitimate integrations. Volume anomalies are the tell: a service account that historically syncs 50 MB/day suddenly pushing 5 GB.
  5. Extortion or sale — driver's license data monetizes well on criminal markets; notification delays compound regulatory exposure under state breach notification statutes.

Exploitation status: Confirmed data theft in the wild — 150,000 records are already out. There is no theoretical component here; the question for every other organization is whether the same pattern is in progress against their own vendor ecosystem right now.

Why traditional controls missed it: Vendor environments sit outside your EDR coverage, your SIEM ingestion, and your identity governance. The defender's leverage is (a) telemetry on your side of the data exchange — API gateways, SFTP endpoints, shared buckets — and (b) contractual/monitoring controls over the vendor. The detections below target the behaviors you can observe.

Detection & Response

The following detections focus on the observable mechanics of bulk PII theft: mass archive creation against data directories, compression-tool abuse by non-admin users, and anomalous outbound transfer volume from service accounts. These fire on behavior, not indicators of compromise that expire in a week.

Sigma Rules

YAML
---
title: Bulk Archive Creation in PII or Data Export Directories
id: 3f8a2c14-7b91-4e56-a2d8-9c1e5f7a3b02
status: experimental
description: Detects compression utilities invoked against directories commonly used for database exports, reports, or shared PII — a hallmark staging behavior in third-party data theft.
references:
  - https://attack.mitre.org/techniques/T1560/001/
author: Security Arsenal
date: 2026/04/06
tags:
  - attack.collection
  - attack.t1560.001
logsource:
  category: process_creation
  product: windows
detection:
  selection_tool:
    Image|endswith:
      - '\7z.exe'
      - '\7za.exe'
      - '\rar.exe'
      - '\winrar.exe'
  selection_path:
    CommandLine|contains:
      - '\exports\'
      - '\reports\'
      - '\data\'
      - '\backup\'
      - '\share\'
  selection_password:
    CommandLine|contains:
      - ' -p'
      - ' -hp'
  condition: selection_tool and selection_path and selection_password
falsepositives:
  - Scheduled backup jobs using password-protected archives — baseline and whitelist known job accounts and paths
level: high
---
title: Non-Administrative User Invoking Archive Utility on Server
id: 8c4d1e67-2a5f-4b38-9d71-6e3b8a2c4f15
status: experimental
description: Detects archive utility execution by accounts outside approved admin/backup roles on servers hosting application or database workloads.
references:
  - https://attack.mitre.org/techniques/T1560/
author: Security Arsenal
date: 2026/04/06
tags:
  - attack.collection
  - attack.t1560
logsource:
  category: process_creation
  product: windows
detection:
  selection:
    Image|endswith:
      - '\7z.exe'
      - '\7za.exe'
      - '\rar.exe'
      - '\tar.exe'
  filter_admins:
    User|contains:
      - 'svc_backup'
      - 'svc_veeam'
      - 'adm_'
  condition: selection and not filter_admins
falsepositives:
  - Developers packaging releases — tune filters per host role
level: medium
---
title: Suspicious Outbound Transfer to Consumer File-Sharing Domain
id: 5e2b9f38-1d47-4c86-b3a9-7f2e6d1a8c44
status: experimental
description: Detects network connections from servers to consumer-grade file sharing or paste services frequently abused for PII exfiltration.
references:
  - https://attack.mitre.org/techniques/T1567/002/
author: Security Arsenal
date: 2026/04/06
tags:
  - attack.exfiltration
  - attack.t1567.002
logsource:
  category: network_connection
  product: windows
detection:
  selection:
    DestinationHostname|contains:
      - 'mega.nz'
      - 'mega.co.nz'
      - 'dropbox.com'
      - 'wetransfer.com'
      - 'file.io'
      - 'anonfiles'
      - 'gofile.io'
      - 'pastebin.com'
  filter_clients:
    Image|endswith:
      - '\chrome.exe'
      - '\msedge.exe'
      - '\firefox.exe'
  condition: selection and not filter_clients
falsepositives:
  - Approved business file transfers — restrict this rule to server VLANs and data-hosting segments for best signal
level: high

KQL — Microsoft Sentinel / Defender Hunt

This query hunts for anomalous outbound data volume from service accounts and servers — the single most reliable signal of bulk PII exfiltration through a vendor or SaaS integration. Baseline deviation is what catches what static rules miss.

KQL — Microsoft Sentinel / Defender
// Hunt: anomalous outbound transfer volume from servers/service accounts (potential PII exfiltration)
let Lookback = 14d;
let Baseline = DeviceNetworkEvents
| where TimeGenerated between (ago(Lookback*2) .. ago(Lookback))
| where InitiatingProcessAccountName has_any ("svc_", "service", "system") or DeviceType =~ "Server"
| summarize AvgDailyBytes = avg(tolong(1)) by DeviceName; // placeholder if byte counts not in table
// Primary: Syslog/CEF-ingested firewall or proxy bytes-out anomaly
CommonSecurityLog
| where TimeGenerated > ago(Lookback)
| where isnotempty(SentBytes) and SentBytes > 0
| summarize TotalSent = sum(SentBytes), Destinations = dcount(DestinationHostName), TopDests = make_set(DestinationHostName, 10) by SourceHostName, SourceUserName, bin(TimeGenerated, 1d)
| where TotalSent > 2147483648  // >2 GB/day from a single source — tune to environment baseline
| order by TotalSent desc
| project TimeGenerated, SourceHostName, SourceUserName, TotalSentGB = round(TotalSent / 1073741824.0, 2), Destinations, TopDests;
// Secondary: mass file access on PII-hosting servers (Defender for Endpoint)
DeviceFileEvents
| where TimeGenerated > ago(1d)
| where FolderPath has_any ("\\exports\\", "\\reports\\", "\\data\\", "\\backup\\")
| where ActionType in ("FileCreated", "FileModified")
| where FileName endswith ".zip" or FileName endswith ".7z" or FileName endswith ".rar"
| summarize ArchiveCount = count(), Accounts = make_set(InitiatingProcessAccountName), Files = make_set(FileName, 20) by DeviceName, bin(TimeGenerated, 1h)
| where ArchiveCount >= 5
| order by ArchiveCount desc;

Velociraptor VQL — Endpoint Hunt

Use this artifact across servers and integration hosts to identify recent staging behavior: archive creation in data directories combined with active outbound connections from non-browser processes.

VQL — Velociraptor
-- Hunt: staged archives in export/data paths plus suspicious outbound connections
LET archives = SELECT FullPath, Size, Mtime
FROM glob(globs=['C:/Data/**/*.zip', 'C:/Data/**/*.7z', 'C:/Exports/**/*.zip', 'C:/Exports/**/*.7z', 'D:/Backups/**/*.rar'])
WHERE Mtime > now() - 604800
ORDER BY Mtime DESC;

LET conns = SELECT Pid, Name, Path, RemoteAddress, RemotePort, Status
FROM netstat()
WHERE Status =~ 'ESTABLISHED'
  AND RemotePort in (443, 22, 21)
  AND NOT Path =~ '(?i)(chrome|msedge|firefox|svchost|teams|onedrive)'
  AND RemoteAddress !~ '^(10\\.|192\\.168\\.|172\\.(1[6-9]|2[0-9]|3[01])\\.)';

SELECT * FROM archives
UNION ALL
SELECT FullPath=NULL, Size=NULL, Mtime=NULL FROM conns LIMIT 1;

SELECT Pid, Name, Path, RemoteAddress, RemotePort, Status FROM conns;

Remediation / Audit Script — Third-Party Data Exposure Verification

Run this against your own environment to enumerate service accounts with broad data access and recent bulk-export activity on database-adjacent hosts — the exact posture that turns a vendor compromise into your breach headline.

PowerShell
# Third-Party Data Exposure Audit - Security Arsenal
# Enumerates service accounts with elevated data access and recent bulk-export artifacts

# 1) Identify service accounts with privileged group membership
Write-Host "[+] Service accounts holding privileged group membership:" -ForegroundColor Cyan
Get-ADGroupMember -Identity "Domain Admins" -Recursive -ErrorAction SilentlyContinue |
  Where-Object { $_.SamAccountName -match 'svc|service|sql|backup|vendor|integration' } |
  Select-Object Name, SamAccountName, ObjectClass |
  Format-Table -AutoSize

Get-ADGroupMember -Identity "Enterprise Admins" -Recursive -ErrorAction SilentlyContinue |
  Where-Object { $_.SamAccountName -match 'svc|service|vendor' } |
  Select-Object Name, SamAccountName |
  Format-Table -AutoSize

# 2) Find service accounts with passwords that never expire (persistence risk if stolen)
Write-Host "[+] Service accounts with non-expiring passwords:" -ForegroundColor Cyan
Get-ADUser -Filter { SamAccountName -like "svc*" -and PasswordNeverExpires -eq $true } -Properties PasswordNeverExpires, PasswordLastSet, LastLogonDate |
  Select-Object SamAccountName, PasswordLastSet, LastLogonDate |
  Format-Table -AutoSize

# 3) Recent bulk archive creation on this host (staging indicator)
Write-Host "[+] Archives created in data/export paths in the last 7 days:" -ForegroundColor Cyan
$cutoff = (Get-Date).AddDays(-7)
$paths = @("C:\Data", "C:\Exports", "C:\Reports", "D:\Backups")
foreach ($p in $paths) {
  if (Test-Path $p) {
    Get-ChildItem -Path $p -Recurse -Include *.zip, *.7z, *.rar -ErrorAction SilentlyContinue |
      Where-Object { $_.LastWriteTime -gt $cutoff -and $_.Length -gt 50MB } |
      Select-Object FullName, @{N='SizeMB';E={[math]::Round($_.Length/1MB,2)}}, LastWriteTime |
      Format-Table -AutoSize
  }
}

# 4) Outbound byte volume per process (top talkers, potential exfil channels)
Write-Host "[+] Top outbound processes by established external connections:" -ForegroundColor Cyan
Get-NetTCPConnection -State Established -ErrorAction SilentlyContinue |
  Where-Object { $_.RemoteAddress -notmatch '^(10\.|192\.168\.|172\.(1[6-9]|2[0-9]|3[01])\.|127\.)' } |
  Group-Object OwningProcess |
  Sort-Object Count -Descending |
  Select-Object -First 15 @{N='PID';E={$_.Name}}, Count,
    @{N='Process';E={(Get-Process -Id $_.Name -ErrorAction SilentlyContinue).ProcessName}} |
  Format-Table -AutoSize

Write-Host "[i] Review output. Any service account in privileged groups, non-expiring vendor credentials, or unexpected large archives warrant immediate investigation." -ForegroundColor Yellow

Remediation: Third-Party Risk Is Your Risk

There is no patch for this breach — the fix is architectural and procedural. Prioritize the following:

  1. Inventory your data egress points now. Enumerate every vendor, SaaS platform, and integration that stores or processes your PII. You cannot protect what you have not mapped. Most organizations discover 30–40% more data-sharing relationships than procurement records show.

  2. Enforce least-privilege on data shared with vendors. The Uber breach exposed driver's license numbers held by a services provider — ask whether that provider needed persistent access to full records, or whether tokenized, time-limited, or on-demand access would have reduced the blast radius.

  3. Contractual telemetry rights. Amend vendor agreements to require security event notification within 24–72 hours, annual independent assessments (SOC 2 Type II minimum), and — critically — the right to receive security logs relevant to your data. Detection time at the vendor directly determines your regulatory notification timeline.

  4. Deploy the detections above. Archive-staging and exfil-volume analytics are technique-based; they catch this class of attack regardless of which vendor is next.

  5. Rotate and vault vendor credentials. Any API keys, SFTP accounts, or service principals shared with third parties should reside in a vault with automatic rotation. Audit the PowerShell output above — non-expiring vendor service account passwords are an unacceptable 2026 posture.

  6. Prepare your notification workflow. With driver's license data exposed, affected individuals face state-mandated notification requirements, and many states require free credit monitoring offers. Run the tabletop: if your vendor calls you Friday at 4 PM, can you notify within statutory deadlines?

  7. For affected individuals: recommend driver's license monitoring where the state DMV supports it, fraud alerts with the credit bureaus, and heightened vigilance against phishing that weaponizes the stolen name/address/license combination.

The lesson from this breach is one we keep relearning at significant cost: your security perimeter extends to every organization that touches your data. Vendor compromise is not an edge case — it is the mainstream initial access vector of 2026.

Related Resources

Security Arsenal Incident Response Services AlertMonitor Platform Book a SOC Assessment incident-response Intel Hub

Is your security operations ready?

Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.