Back to Intelligence

FBI NCIC Breach: Detecting and Responding to Compromise of National Law Enforcement Databases

SA
Security Arsenal Team
November 22, 2026
13 min read

Federal systems operated by the FBI have reportedly been compromised for the second time in roughly two months — and this time the target is the National Crime Information Center (NCIC), the centralized index that local, state, tribal, and federal law enforcement agencies query millions of times per day. NCIC isn't a bureau-internal system; it's the shared circulatory system of U.S. criminal justice data — wanted persons, stolen vehicles, protective orders, gang files, sex offender registries, and officer-safety flagging.

Let me be direct about why this matters beyond Washington: every agency that touches NCIC is now in the blast radius. If an adversary holds valid credentials or session access into the NCIC ecosystem, the exposure isn't limited to FBI data — it extends to every local terminal, every CJIS-connected workstation in a patrol car, and every state fusion center that trusts those pipes. Twenty-one interconnected systems and roughly a dozen discrete data categories sit behind that front door.

I've run IR engagements where a shared inter-agency platform became the pivot point for lateral movement into dozens of downstream tenants. The pattern is always the same: the platform operator gets breached, and the tenants discover they have no independent visibility into whether their own queries, credentials, or terminals were abused. Don't be that tenant.

Technical Analysis: What We Know and What We Should Assume

Affected Systems and Trust Boundaries

ComponentExposure
NCICCentral index for criminal justice data sharing across local, state, tribal, and federal agencies
CJIS-adjacent infrastructureTerminal agency coordinators, state control terminal agencies (CTAs), and mobile data terminals that proxy queries upstream
Interconnected state/local systemsAny agency with automated NCIC query interfaces or stored credentials for programmatic access

Likely Attack Chain (Defender's Assessment)

No CVE has been disclosed in connection with this breach, and we should not speculate on a specific software flaw. Based on the reported pattern — repeated access to a federal law enforcement platform over a compressed timeframe — the realistic intrusion hypotheses, in descending order of likelihood, are:

  1. Credential compromise of a privileged or federated account — phishing, infostealer logs, or token theft yielding valid access without exploiting a software vulnerability at all.
  2. Abuse of an externally exposed authentication surface — VPN, remote access gateway, or a web-facing portal with weak or missing MFA enforcement.
  3. Downstream pivot — compromise of a state-level or local agency with trusted NCIC connectivity, used as a beachhead to reach the federal platform.

This is the uncomfortable truth defenders need to internalize: most breaches of high-value government databases in 2025–2026 are identity attacks, not exploit chains. A stolen credential that passes MFA-fatigue or sits on an MFA-exempt service account bypasses every WAF and EDR control you've deployed.

Exploitation Status

  • Confirmed active intrusion: Yes — reported breach of FBI systems, second incident in approximately two months.
  • CVE / PoC: None disclosed. No specific vulnerability identifier is associated with this incident at time of writing.
  • CISA KEV: Not applicable — no CVE attributed.
  • Attribution: Unconfirmed at time of writing. Treat as an active, resourced adversary with persistent access intent.

The Data at Stake

NCIC query access exposes wanted-person files, protection orders, supervised release data, stolen property records, and — critically — officer-safety information. In the hands of a criminal enterprise, this enables counter-surveillance (checking whether a subject is flagged before a meet), witness intimidation, and operational compromise of undercover activity. In the hands of a nation-state, it's a targeting and recruitment dataset. Either way, query logs themselves are sensitive — they reveal investigative interest and active case geography.

Detection & Response

The detections below are built for the tenant side — your agency's side of the NCIC trust relationship — because that's where most of my readers sit. You cannot instrument the FBI's infrastructure, but you absolutely can instrument your own terminals, query gateways, and identity providers.

Sigma Rules

These rules target the two highest-fidelity observable behaviors: access to criminal justice portals from non-allowlisted sources, and anomalous authenticated access patterns to CJIS-connected infrastructure.

YAML
---
title: NCIC or CJIS Portal Access From Non-Allowlisted Source
title_note: Detects web access to criminal justice information portals from source IPs outside approved agency ranges.
id: 8f4c2b1a-6d3e-4a9f-b5c7-2e1d0a3f9b8c
status: experimental
description: Identifies HTTP(S) requests to NCIC/CJIS-facing portals originating from source addresses outside the agency's approved terminal ranges. A successful breach of a federated law enforcement database is frequently preceded or accompanied by access from infrastructure that has no legitimate reason to query criminal justice data.
references:
  - https://www.bleepingcomputer.com/news/security/fbi-reportedly-hacked-again-this-time-targeting-nationwide-police-database/
  - https://attack.mitre.org/techniques/T1190/
  - https://attack.mitre.org/techniques/T1078/
author: Security Arsenal
date: 2026/01/19
tags:
  - attack.initial_access
  - attack.t1190
  - attack.t1078
logsource:
  category: proxy
  product: proxy
detection:
  selection_target:
    c-uri|contains:
      - 'ncic'
      - 'cjis'
      - 'nlets'
      - 'leoweb'
  filter_internal:
    src_ip|cidr:
      - '10.0.0.0/8'
      - '172.16.0.0/12'
      - '192.168.0.0/16'
  condition: selection_target and not filter_internal
falsepositives:
  - Mobile data terminals (MDTs) routing through carrier NAT ranges - baseline these ranges and add them to the filter
  - Vendors performing authorized portal testing - require change tickets
level: high
---
title: Excessive Authentication Failures Followed By Success On CJIS-Connected Host
title_note: Detects password spray or credential stuffing behavior against systems with access to criminal justice databases.
id: 3a9d7e52-1b4c-4f68-9d2a-7c5e0b8f3a1d
status: experimental
description: Correlates a burst of failed logons (Type 3 network or Type 10 remote interactive) against a host followed by a successful logon from the same source within a short window. Compromised credentials are the leading intrusion vector against government data-sharing platforms, and brute validation of stolen credentials generates this signature on the destination host.
references:
  - https://www.bleepingcomputer.com/news/security/fbi-reportedly-hacked-again-this-time-targeting-nationwide-police-database/
  - https://attack.mitre.org/techniques/T1110/
author: Security Arsenal
date: 2026/01/19
tags:
  - attack.credential_access
  - attack.t1110
logsource:
  category: authentication
  product: windows
detection:
  selection_fail:
    LogonType:
      - 3
      - 10
    EventType: 'failure'
  timeframe: 10m
  condition: selection_fail | count() by WorkstationName, SourceAddress > 20
falsepositives:
  - Service accounts with expired cached credentials across terminal fleets - tune per host after baselining
  - RADIUS/NPS misconfiguration generating retry storms
level: high

A note on tuning: the first rule will fire on day one if you haven't mapped your MDT carrier ranges. That's fine — treat the initial deployment as an asset-discovery exercise. Every agency I've assessed that runs this rule finds at least one terminal it didn't know about.

KQL — Microsoft Sentinel / Defender

This hunt query looks for the breach behavior that matters most post-compromise: anomalous query volume or new-source access against criminal justice data infrastructure, ingested via proxy/firewall logs (CommonSecurityLog) or Syslog. It flags sources that have never historically touched your NCIC/CJIS endpoints, or accounts whose query rate spikes well beyond their baseline.

KQL — Microsoft Sentinel / Defender
// Hunt: Anomalous access to NCIC/CJIS/NLETS endpoints - new sources and volume spikes
let cjisEndpoints = dynamic(["ncic", "cjis", "nlets", "leoweb"]);
let lookback = 14d;
let window = 1d;
let baseline =
    CommonSecurityLog
    | where TimeGenerated between (ago(lookback) .. ago(window))
    | where tolower(RequestURL) has_any (cjisEndpoints)
    | summarize BaselineSources = make_set(SourceIP), BaselineUsers = make_set(RequestUserName);
CommonSecurityLog
| where TimeGenerated > ago(window)
| where tolower(RequestURL) has_any (cjisEndpoints)
| extend KnownSource = SourceIP in (toscalar(baseline | project BaselineSources))
| summarize
    QueryCount = count(),
    FirstSeen = min(TimeGenerated),
    LastSeen = max(TimeGenerated),
    DistinctUsers = dcount(RequestUserName),
    Users = make_set(RequestUserName, 20)
    by SourceIP, RequestURL, KnownSource
| where KnownSource == false or QueryCount > 500
| order by QueryCount desc

Two thresholds worth your attention: QueryCount > 500 per source per day is a starting point for a mid-size agency — set it from your own 90-day p99. And DistinctUsers per source IP above 2 or 3 almost always indicates a shared workstation or a proxied query gateway; anything above that from a new source is a paging event.

Velociraptor VQL

For endpoint triage on a terminal or query gateway you suspect was used as the pivot, hunt for live connections to law enforcement data infrastructure alongside recently written bulk-export artifacts — CSV/TXT dumps are the canonical output of NCIC harvesting.

VQL — Velociraptor
-- Hunt: NCIC/CJIS access from this endpoint plus suspicious bulk data staging
-- Run on query gateways, dispatch workstations, and MDT hosts under suspicion.

LET connections = SELECT Pid, Name, RemoteAddress, RemotePort, Status
FROM netstat()
WHERE RemoteAddress =~ 'ncic|cjis|nlets'
   OR Name =~ 'ncic|cjis|leoweb'

LET staged_exports = SELECT FullPath, Size, Mtime, Btime
FROM glob(globs=[
    'C:/Users/*/Downloads/*.csv',
    'C:/Users/*/Desktop/*.csv',
    'C:/Temp/*.csv',
    'C:/ProgramData/*/*.csv'
])
WHERE Size > 100000
  AND Mtime > now() - 86400 * 7

SELECT 'network' AS ArtifactType, Pid AS Pid, Name AS Name,
       RemoteAddress AS Detail, Status AS Status
FROM connections
UNION ALL
SELECT 'staged_file' AS ArtifactType, NULL AS Pid, FullPath AS Name,
       format(format='%d bytes, modified %s', args=[Size, Mtime]) AS Detail,
       NULL AS Status
FROM staged_exports

This is deliberately narrow. Bulk CSV staging in user-writable paths on a CJIS-connected host, co-occurring with a live session to criminal justice infrastructure, is the exfiltration sequence I'd want a hunter to find before the adversary archives and egresses it.

Remediation Script — PowerShell

Run this on CJIS-connected Windows terminals and query gateways to audit authentication hygiene and surface the logon anomalies that identity-based intrusions leave behind.

PowerShell
# NCIC/CJIS Host Posture Audit - run as Administrator on CJIS-connected systems
# Security Arsenal IR toolkit - review output before sharing externally

$report = @()
$since = (Get-Date).AddDays(-14)

# 1. Remote interactive (RDP/network) logons - source IP distribution
Write-Host "[*] Collecting Type 3/10 successful logons (last 14 days)..." -ForegroundColor Cyan
$logons = Get-WinEvent -FilterHashtable @{LogName='Security'; Id=4624; StartTime=$since} -ErrorAction SilentlyContinue |
  Where-Object { $_.Properties[8].Value -in @(3,10) } |
  ForEach-Object {
    [PSCustomObject]@{
      TimeCreated  = $_.TimeCreated
      Account      = $_.Properties[5].Value
      LogonType    = $_.Properties[8].Value
      SourceIP     = $_.Properties[18].Value
      Workstation  = $_.Properties[11].Value
    }
  }

$external = $logons | Where-Object {
  $_.SourceIP -and
  $_.SourceIP -notmatch '^(10\.|192\.168\.|172\.(1[6-9]|2[0-9]|3[01])\.|127\.|::1|-)'
}
if ($external) {
  Write-Host "[!] EXTERNAL source IPs with successful logons:" -ForegroundColor Red
  $external | Group-Object SourceIP | Sort-Object Count -Descending | Format-Table Count, Name
  $report += $external
}

# 2. Privileged accounts with stale passwords (CJIS requires 90-day rotation for standard users, stricter for privileged)
Write-Host "[*] Checking privileged account password age..." -ForegroundColor Cyan
$admins = Get-LocalGroupMember -Group 'Administrators' -ErrorAction SilentlyContinue
foreach ($m in $admins) {
  if ($m.ObjectClass -eq 'User') {
    $u = Get-LocalUser -Name $m.Name.Split('\')[-1] -ErrorAction SilentlyContinue
    if ($u -and $u.PasswordLastSet -and $u.PasswordLastSet -lt (Get-Date).AddDays(-60)) {
      Write-Host "[!] Privileged account '$($u.Name)' password last set: $($u.PasswordLastSet)" -ForegroundColor Yellow
    }
  }
}

# 3. Verify CJIS-mandated audit policy is actually enabled
Write-Host "[*] Auditing logon/logoff audit policy state..." -ForegroundColor Cyan
auditpol /get /subcategory:"Logon","Special Logon","Other Logon/Logoff Events" | Out-String | Write-Host

# 4. Flag any service accounts with interactive logon capability (should be none on terminals)
Write-Host "[*] Checking for services running as user accounts..." -ForegroundColor Cyan
Get-CimInstance Win32_Service |
  Where-Object { $_.StartName -and $_.StartName -notmatch 'LocalSystem|NetworkService|LocalService|NT AUTHORITY|NT SERVICE' } |
  Select-Object Name, StartName, State | Format-Table

if ($report) {
  $report | Export-Csv -Path ".\ncic_posture_audit_$(Get-Date -Format yyyyMMdd).csv" -NoTypeInformation
  Write-Host "[*] Findings exported to .\ncic_posture_audit_$(Get-Date -Format yyyyMMdd).csv" -ForegroundColor Green
}
Write-Host "[*] Audit complete. Any external source IPs above require immediate credential reset and CJIS Security Office notification." -ForegroundColor Cyan

Remediation and Hardening

There is no patch for a credential-driven breach. Remediation here is architectural and procedural — and for any agency with NCIC access, it's anchored in the FBI CJIS Security Policy, which is contractually binding on every terminal agency.

Immediate Actions (0–72 hours)

  1. Rotate every credential that touches NCIC-adjacent infrastructure — terminal operator accounts, programmatic query service accounts, VPN credentials for CJIS-connected segments. Assume any credential used in the last 90 days is in an adversary's possession.
  2. Invalidate all active sessions and tokens against your IdP for accounts with criminal justice data access. Password resets without session revocation leave stolen session cookies valid.
  3. Pull 90 days of NCIC transaction logs for your ORI and diff query volume against baseline. Look for after-hours query patterns, bulk sequential queries (indicative of harvesting), and queries from terminal IDs you don't recognize.
  4. Contact your state CJIS Systems Officer (CSO) and the CJIS ISO if you identify any anomaly — breach notification obligations under CJIS policy and state statute have hard clocks.

Short-Term Hardening (1–4 weeks)

  1. Enforce phishing-resistant MFA — FIDO2/passkeys or PIV/CAC — for every account with CJIS-connected system access. SMS and push-based MFA are demonstrably insufficient against the adversary class targeting federal platforms in 2026. The CJIS Security Policy's advanced authentication requirement (Section 5.6.2.2) is the floor, not the ceiling.
  2. Segment CJIS-connected terminals onto a dedicated VLAN with egress filtering permitting only required NCIC/NLETS/state-switch destinations. A patrol MDT should not be able to reach the internet, full stop.
  3. Deploy the detections above and route alerts to a human within your defined SLA. A Sigma rule that pages into an unmanaged mailbox is a control on paper only.
  4. Audit your automated query interfaces — message-switch integrations, RMS/CAD middleware with stored NCIC credentials. These service accounts are typically MFA-exempt, over-privileged, and unmonitored. They are your highest-risk identity.

Structural Fixes (This Quarter)

  1. Move to just-in-time privileged access for CJIS administration. Standing administrative rights on criminal justice infrastructure is what turns a phish into a breach.
  2. Execute a tabletop exercise on exactly this scenario: federated platform compromise, your ORI's query logs in adversary hands, officer-safety data exposed. If your IR plan doesn't cover "the platform we trust got breached," it doesn't cover 2026.
  3. Review CJIS Security Policy v5.9+ compliance against your actual configuration, not your attestation. In every engagement where I've tested a CJIS-attested environment, the gap between the paperwork and the running config was the finding.

What Federal Tenants Should Demand

If you're a CISO at a downstream agency, you have limited leverage over federal remediation — but you have more than you think. Demand: (a) formal notification scope — which data categories, which ORIs, what timeframe; (b) IOCs and TTPs from the federal IR, even under TLP:AMBER; (c) confirmation of whether federated authentication tokens or certificates were in scope. Silence is not an acceptable answer when your officers' safety data is on the table.

The Bigger Lesson

Two breaches of the same federal law enforcement infrastructure in two months tells us the adversary achieved persistence, not just access. Every organization that operates a shared trust platform — or depends on one — should treat this as a live-fire drill for their own environment. Identity is the perimeter, query logs are evidence, and the platform you trust is an extension of your attack surface whether you manage it or not.

Related Resources

Security Arsenal Managed SOC Services AlertMonitor Platform Book a SOC Assessment soc-mdr Intel Hub

Is your security operations ready?

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