Back to Intelligence

BAE Systems Breach Disrupts Federal Bankruptcy Court Noticing — Third-Party Compromise Response Guide

SA
Security Arsenal Team
October 17, 2026
12 min read

When a single contractor outage can silence an entire branch of the federal judiciary, third-party risk stops being a compliance checkbox and becomes an operational resilience problem. That is exactly what happened when the Administrative Office of the US Courts disclosed that a cyberattack on government contractor BAE Systems took down the noticing system used by federal bankruptcy courts.

What Happened

According to reporting by BleepingComputer, the Administrative Office of the US Courts (AO) confirmed that a cyberattack against BAE Systems — the contractor operating the bankruptcy court noticing infrastructure — disrupted the system responsible for sending court notices to creditors, attorneys, trustees, and other parties in bankruptcy proceedings. The Bankruptcy Noticing Center (BNC) and related noticing functions are the connective tissue of the bankruptcy process: when they go dark, legally mandated notifications stop flowing, deadlines become legally ambiguous, and thousands of pending cases are effectively frozen in procedural limbo.

Key facts defenders should internalize:

  • The victim of record is a contractor, not the judiciary itself. The federal courts' case management systems (CM/ECF, PACER) were not reported as directly compromised — the operational impact cascaded from a third party.
  • Impact is service disruption, not (publicly confirmed) data theft. As of this writing, no CVE, malware family, or threat actor attribution has been disclosed. The intrusion vector, dwell time, and scope of data exposure remain under investigation.
  • This follows a pattern. The judiciary has been a recurring target — the AO previously disclosed a serious breach of its case management system attributed to nation-state actors. Adversaries understand that legal infrastructure holds extraordinarily sensitive sealed records and that disrupting it creates outsized pressure.

Why Defenders Need to Act — Even If You're Not BAE Systems

This incident is a textbook supply-chain / third-party compromise scenario, and the defensive lessons generalize to every organization that outsources a business-critical function:

  1. Your vendor's compromise is your outage. The courts did nothing wrong at the perimeter of their own network and still lost a mission-critical service. If you have not mapped which vendors hold which operational chokepoints, you have a blind spot the size of this incident.
  2. Contractors handling government data are high-value targets. Defense industrial base (DIB) contractors and government service providers are persistently targeted by nation-state and criminal actors precisely because they aggregate access to many downstream customers.
  3. The initial access vector matters less than the blast radius. Whether this was a phishing-led credential compromise, an exploited edge appliance, or a ransomware intrusion, the detective controls that catch contractor-environment intrusions early are the same — and they are the difference between an outage announcement and a contained event.

Because no CVE or specific malware indicators have been publicly attributed to this incident, this post focuses on the behavioral detection layer — the TTPs that characterize intrusions into contractor and service-provider environments before they become customer-facing outages.

Technical Analysis

Attack Pattern: Third-Party Service Provider Intrusion

While the specific initial vector in the BAE Systems incident has not been disclosed, intrusions against government contractors that result in production service disruption typically follow a recognizable chain:

  1. Initial access — spear-phishing against contractor employees, exploitation of internet-facing remote access (VPN, VDI, remote monitoring and management tooling), or abused third-party credentials.
  2. Persistence and privilege escalation — web shells on edge systems, creation of rogue accounts, or abuse of legitimate RMM platforms already deployed in the environment.
  3. Lateral movement toward the hosted service — pivoting from the corporate segment into the application/hosting tier that operates the customer-facing system (here, the noticing infrastructure).
  4. Impact — encryption, destructive actions, or forced defensive shutdown of the hosted application, which is what manifests publicly as 'the system is down.'

Why Behavioral Detection Wins Here

No IOC list has been published for this incident. That means detections built on hashes or IPs would be fiction. What is reliable:

  • Anomalous service-account behavior against the hosted application tier (new logon sources, impossible travel, interactive logons where only service logons are expected).
  • RMM and remote access tooling abuse — tools like AnyDesk, ScreenConnect, and TeamViewer remain the most common lateral movement and persistence vehicles in contractor-environment intrusions, and unauthorized instances are high-signal.
  • Mass service disruption precursors — mass process termination, backup deletion (VSS destruction), and mass file modification are the observable last mile before an outage.

The detections below target exactly those behaviors.

Detection & Response

Sigma Rules

YAML
---
title: Unauthorized Remote Access Tool Execution in Contractor Environment
id: 8c4f2a91-3d7b-4e6a-9f21-5b8d0c3e7a44
status: experimental
description: Detects execution of commonly abused remote access tools (AnyDesk, ScreenConnect, TeamViewer, Atera, Splashtop) from unexpected paths or by non-admin users. RMM abuse is the dominant persistence/lateral movement mechanism in third-party contractor intrusions such as the BAE Systems incident that disrupted federal bankruptcy court noticing.
references:
  - https://www.bleepingcomputer.com/news/security/bae-systems-cyberattack-disrupts-us-federal-court-operations/
  - https://attack.mitre.org/techniques/T1219/
author: Security Arsenal
date: 2026/02/18
tags:
  - attack.command_and_control
  - attack.t1219
logsource:
  category: process_creation
  product: windows
detection:
  selection_tool:
    Image|endswith:
      - '\anydesk.exe'
      - '\screenconnect.client.exe'
      - '\teamviewer.exe'
      - '\teamviewer_service.exe'
      - '\atera_agent.exe'
      - '\splashtop_streamer.exe'
  filter_known_paths:
    Image|startswith:
      - 'C:\Program Files\AnyDesk'
      - 'C:\Program Files\TeamViewer'
      - 'C:\Program Files (x86)\TeamViewer'
      - 'C:\Program Files (x86)\ScreenConnect Client'
  condition: selection_tool and not filter_known_paths
falsepositives:
  - IT support staff running portable RMM binaries during sanctioned remote sessions; maintain an allowlist of approved RMM paths and update the filter
level: high
---
title: Volume Shadow Copy Deletion Prior to Service Disruption
id: 2e7b9c14-6a1f-4d83-b5e2-8f0a4c9d6e31
status: experimental
description: Detects deletion of volume shadow copies or backup catalog manipulation via vssadmin, wmic, bcdedit, or wbadmin. This is a high-fidelity precursor to destructive/encryption impact that typically precedes the kind of forced service outage seen in the federal bankruptcy court noticing disruption.
references:
  - https://www.bleepingcomputer.com/news/security/bae-systems-cyberattack-disrupts-us-federal-court-operations/
  - https://attack.mitre.org/techniques/T1490/
author: Security Arsenal
date: 2026/02/18
tags:
  - attack.impact
  - attack.t1490
logsource:
  category: process_creation
  product: windows
detection:
  selection_vssadmin:
    Image|endswith: '\vssadmin.exe'
    CommandLine|contains:
      - 'delete shadows'
      - 'Delete Shadows'
  selection_wmic:
    Image|endswith: '\wmic.exe'
    CommandLine|contains: 'shadowcopy delete'
  selection_bcdedit:
    Image|endswith: '\bcdedit.exe'
    CommandLine|contains: 'recoveryenabled no'
  selection_wbadmin:
    Image|endswith: '\wbadmin.exe'
    CommandLine|contains:
      - 'delete catalog'
      - 'delete systemstatebackup'
  condition: 1 of selection_*
falsepositives:
  - Rare; legitimate backup administrators occasionally clear shadow copies during storage reclamation — correlate with change tickets
level: critical
---
title: Interactive Logon by Application Service Account
id: 5a1d8e36-9c4b-4f27-a3d8-2e6b1f7c9052
status: experimental
description: Detects interactive (Type 2) or RemoteInteractive (Type 10) logons by accounts matching common service account naming patterns. Service accounts backing hosted applications (such as the court noticing system) should never log on interactively; this behavior is a strong indicator of credential theft and hands-on-keyboard intrusion in a service provider environment.
references:
  - https://www.bleepingcomputer.com/news/security/bae-systems-cyberattack-disrupts-us-federal-court-operations/
  - https://attack.mitre.org/techniques/T1078/
author: Security Arsenal
date: 2026/02/18
tags:
  - attack.persistence
  - attack.initial_access
  - attack.t1078
logsource:
  category: authentication
  product: windows
detection:
  selection:
    LogonType:
      - 2
      - 10
    TargetUserName|contains:
      - 'svc-'
      - 'svc_'
      - '-svc'
      - 'service'
  condition: selection
falsepositives:
  - Poorly governed environments where admins interactively log on with service accounts; tune TargetUserName patterns to your actual naming convention
level: high

KQL — Microsoft Sentinel / Defender

The following hunt queries target the two most actionable behaviors: rogue RMM presence across the contractor estate, and backup/shadow-copy destruction preceding service impact.

KQL — Microsoft Sentinel / Defender
// Hunt 1: Unauthorized remote access tooling across endpoints (Defender)
// Baseline your approved RMM list first; anything outside it in a contractor
// or hosted-service environment warrants immediate investigation.
DeviceProcessEvents
| where TimeGenerated > ago(14d)
| where FileName has_any ("anydesk.exe", "screenconnect.client.exe", "teamviewer.exe",
    "teamviewer_service.exe", "atera_agent.exe", "splashtop_streamer.exe", "rustdesk.exe", "ultraviewer.exe")
| extend RemoteIP = tostring(parse_json(InitiatingProcessCommandLine))
| summarize FirstSeen = min(TimeGenerated), LastSeen = max(TimeGenerated),
    LaunchCount = count(), DistinctUsers = dcount(AccountName),
    Users = make_set(AccountName), CommandLines = make_set(ProcessCommandLine)
    by DeviceName, FileName, FolderPath
| order by FirstSeen asc

// Hunt 2: Shadow copy deletion / backup tampering preceding service disruption (Sentinel, SecurityEvent 4688)
SecurityEvent
| where TimeGenerated > ago(14d)
| where EventID == 4688
| where NewProcessName has_any ("vssadmin.exe", "wmic.exe", "wbadmin.exe", "bcdedit.exe")
| where CommandLine has_any ("delete shadows", "shadowcopy delete", "delete catalog",
    "recoveryenabled no", "delete systemstatebackup")
| project TimeGenerated, Computer, Account, NewProcessName, CommandLine, ParentProcessName
| order by TimeGenerated desc

// Hunt 3: Interactive logons by service accounts into the hosted application tier (Sentinel)
SecurityEvent
| where TimeGenerated > ago(7d)
| where EventID == 4624
| where LogonType in (2, 10)
| where TargetUserName matches regex @"(?i)(svc[-_]|[-_]svc|service)"
| summarize LogonCount = count(), SourceIPs = make_set(IpAddress), Hosts = make_set(Computer)
    by TargetUserName, LogonType
| order by LogonCount desc

Velociraptor VQL

Use this artifact to sweep a contractor or hosted-service environment for unauthorized RMM binaries and recently dropped executables in user-writable paths — the most common staging pattern before an intrusion reaches the application tier.

VQL — Velociraptor
-- Hunt for unauthorized RMM binaries and recent suspicious executables
-- Deploy across the estate as a hunt; scope to servers hosting customer-facing services first
LET rmm_paths = SELECT FullPath, Size, Mtime
FROM glob(globs=[
    'C:/Users/*/Downloads/*anydesk*',
    'C:/Users/*/Downloads/*teamviewer*',
    'C:/Users/*/Downloads/*screenconnect*',
    'C:/Users/*/Downloads/*rustdesk*',
    'C:/Users/*/AppData/**/anydesk.exe',
    'C:/ProgramData/**/screenconnect*',
    'C:/ProgramData/**/atera*',
    'C:/Temp/*.exe'
])

LET suspect_procs = SELECT Pid, Name, Exe, CommandLine, Username, CreateTime
FROM pslist()
WHERE Name =~ '(?i)(anydesk|teamviewer|screenconnect|rustdesk|ultraviewer|atera|splashtop)'
   OR CommandLine =~ '(?i)(vssadmin.*delete|shadowcopy.*delete|bcdedit.*recoveryenabled)'

SELECT * FROM rmm_paths
UNION ALL
SELECT Exe AS FullPath, NULL AS Size, CreateTime AS Mtime FROM suspect_procs

Remediation & Verification Script

This PowerShell script audits a Windows estate (run via your EDR/remote shell of choice) for unauthorized RMM tooling, recent shadow copy deletion, and interactive service-account logons — the three behaviors above — and exports findings for triage.

PowerShell
# Security Arsenal — Third-Party Intrusion Triage Script
# Run with elevated privileges on servers hosting critical services
# Outputs a consolidated CSV of findings per host

$Findings = @()

# 1. Scan for unauthorized RMM binaries outside approved paths
$ApprovedPaths = @("C:\Program Files\AnyDesk", "C:\Program Files\TeamViewer", "C:\Program Files (x86)\TeamViewer")
$RmmNames = @("anydesk.exe","teamviewer.exe","teamviewer_service.exe","screenconnect.client.exe","atera_agent.exe","splashtop_streamer.exe","rustdesk.exe")

foreach ($Name in $RmmNames) {
    $Hits = Get-CimInstance Win32_Process -Filter "Name='$Name'" -ErrorAction SilentlyContinue
    foreach ($Hit in $Hits) {
        $ExePath = $Hit.ExecutablePath
        $Approved = $false
        foreach ($Path in $ApprovedPaths) { if ($ExePath -like "$Path*") { $Approved = $true } }
        if (-not $Approved) {
            $Findings += [PSCustomObject]@{
                Category = "UnauthorizedRMM"
                Detail   = "$Name running from $ExePath (PID $($Hit.ProcessId))"
                Severity = "High"
                Host     = $env:COMPUTERNAME
            }
        }
    }
}

# 2. Check shadow copy status — missing shadows on a server is an impact red flag
$Shadows = Get-CimInstance Win32_ShadowCopy -ErrorAction SilentlyContinue
if (-not $Shadows) {
    $Findings += [PSCustomObject]@{
        Category = "ShadowCopyAbsent"
        Detail   = "No volume shadow copies present — investigate vssadmin/wmic deletion"
        Severity = "Critical"
        Host     = $env:COMPUTERNAME
    }
}

# 3. Hunt security log for vssadmin/wmic/bcdedit tampering (last 14 days)
$TamperEvents = Get-WinEvent -FilterHashtable @{LogName='Security'; Id=4688; StartTime=(Get-Date).AddDays(-14)} -ErrorAction SilentlyContinue |
    Where-Object { $_.Message -match 'vssadmin.*delete shadows|shadowcopy delete|bcdedit.*recoveryenabled|wbadmin.*delete catalog' }
foreach ($Evt in $TamperEvents) {
    $Findings += [PSCustomObject]@{
        Category = "BackupTampering"
        Detail   = ($Evt.Message -split "`n" | Select-String 'Command Line').ToString()
        Severity = "Critical"
        Host     = $env:COMPUTERNAME
    }
}

# 4. Interactive logons by service accounts (last 7 days)
$SvcLogons = Get-WinEvent -FilterHashtable @{LogName='Security'; Id=4624; StartTime=(Get-Date).AddDays(-7)} -ErrorAction SilentlyContinue |
    Where-Object { $_.Message -match 'Logon Type:\s+(2|10)' -and $_.Message -match '(?i)svc[-_]|[-_]svc' }
foreach ($Evt in $SvcLogons) {
    $Findings += [PSCustomObject]@{
        Category = "SvcAcctInteractiveLogon"
        Detail   = "Service account interactive/RDP logon at $($Evt.TimeCreated)"
        Severity = "High"
        Host     = $env:COMPUTERNAME
    }
}

$Findings | Export-Csv -Path "C:\Temp\intrusion_triage_$env:COMPUTERNAME.csv" -NoTypeInformation
Write-Host "[+] Triage complete. $($Findings.Count) findings exported to C:\Temp\intrusion_triage_$env:COMPUTERNAME.csv"

Remediation

Since no CVE or specific exploited product has been disclosed in this incident, remediation is architectural rather than patch-specific. Prioritize in this order:

Immediate (24–48 hours):

  1. Inventory and baseline all remote access tooling. Every RMM binary outside an approved, version-pinned allowlist is an incident until proven otherwise. Enforce application control (AppLocker/WDAC) blocking unapproved RMM executables on servers hosting critical services.
  2. Audit service account hygiene on hosted application tiers. Disable interactive logon rights for all service accounts via GPO (Deny log on locally, Deny log on through Remote Desktop Services). Alert on any violation.
  3. Verify backup integrity and isolation. Confirm shadow copies exist, backups are immutable or offline, and backup administrative credentials are tiered away from the production environment they protect.
  4. Force credential rotation for any vendor with access to your environment. If your organization consumes services from a compromised provider, treat shared credentials, API keys, and federation trust as potentially exposed until the provider confirms scope.

Near-term (1–4 weeks):

  1. Map operational dependencies on third parties. For every vendor, document: what service they operate, what data they hold, what your outage looks like if they go down, and what your contractual notification SLA is. The federal courts learned the answer to the third question in public.
  2. Segment vendor connectivity. Vendor access should traverse dedicated, monitored, just-in-time pathways — never standing VPN access into flat network segments.
  3. Test your continuity plan for vendor-caused outages. The AO's courts had to fall back to manual noticing workflows. If your equivalent answer is 'we would improvise,' that is a finding.

Strategic:

  1. Contractual security requirements with teeth. Incident notification timelines (24–72 hours), forensic cooperation clauses, evidence preservation obligations, and minimum control baselines (NIST SP 800-171 for government contractors) must be enforceable, audited, and exercised — not just attested annually.

The Bottom Line

The BAE Systems intrusion did not need a novel zero-day to paralyze a federal function — it needed a contractor with access and an adversary with patience. The detections above target the universal mechanics of these intrusions: rogue remote access tooling, backup destruction ahead of impact, and service accounts behaving like humans. Deploy them before your organization's name is in the next headline about someone else's breach.

Related Resources

Security Arsenal Healthcare Cybersecurity AlertMonitor Platform Book a SOC Assessment healthcare Intel Hub

Is your security operations ready?

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