← Back to Feed CACHED · 2026-10-07 05:31:28 · CACHE_KEY CVE-2026-17583
CVE-2026-17583 · CWE-353 · Disclosed 2026-08-05

Thermo Fisher Applied Biosystems Genetic Analyzer Missing Integrity Check on .fsa/.hid Output Files

ASSESSED — NOISGATE
Vendor
—
—
—
Reassessed
—
—
—
Verdict: —
Do you agree?
01 · The Real Story

Someone can rewrite the crime-lab's DNA results with Notepad and nobody's instrument will blink

CVE-2026-17583 is a missing-integrity-check vulnerability in Thermo Fisher Applied Biosystems Genetic Analyzers. The .fsa and .hid output files — the raw digital representation of capillary-electrophoresis DNA fragment analysis — carry no cryptographic signature or checksum. Any user or process that can write to the file share where results land can silently alter allele calls, peak heights, or even splice two individuals' profiles into a single file. Affected products span the 3500/3500xL (≤4.0.2), 3730/3730xL (≤5.0.2), SeqStudio (≤1.2.5), SeqStudio Flex (≤1.2.0), and GeneMapper ID-X (≤1.7.3). Three legacy platforms — the 3130, 3100/3100-Avant, and 310 — are end-of-life and will never receive a fix. The flaw has existed in every version of the file format since approximately 1995.

Thermo Fisher and CISA rate this HIGH at 8.4 (CVSS 3.1) / 8.2 (CVSS 4.0), and that assessment is directionally correct. The CVSS 3.1 vector scores C:H/I:H/A:H, which overstates confidentiality and availability impact — this is purely an *integrity* vulnerability; the 4.0 vector (VI:H, SI:H, VC/VA/SC/SA all N) is more accurate. The local-access requirement (AV:L) and the extremely narrow installed base (forensic and clinical DNA labs, not general IT) mean the population of reachable targets is small and the attacker needs either insider access or a prior compromise of laboratory infrastructure. However, the *consequence* of successful exploitation — fabricated or destroyed forensic evidence, misdiagnosed patients — is disproportionately severe compared to a typical 8.4-rated IT vulnerability. noisgate downgrades the score to 7.5 to reflect the friction but maintains the HIGH floor because nearly 100% of deployments sit in safety-critical forensic or clinical roles.

"Safety-critical integrity flaw in forensic DNA equipment — HIGH floor holds despite local-only access."
02 · The Attack Path

5 steps from start to impact.

STEP 01

Gain access to the lab network or workstation

The attacker must reach the Windows workstation or file share where .fsa/.hid files are stored. In forensic labs this is typically an isolated VLAN or air-gapped subnet. Access paths include insider credentials (lab technician, IT admin), a compromised lab workstation via phishing or USB, or a misconfigured network share bridging the lab to the corporate LAN. There is no remote network service exposed by the instrument software itself.
Conditions required:
  • Local or network access to the file storage location
  • Valid credentials or prior host compromise on the lab network
Where this breaks in practice:
  • Forensic labs typically operate on isolated networks per accreditation requirements (ISO 17025, FBI QAS)
  • Physical access controls (badge, biometric) limit casual insider scenarios
  • Lab IT is usually managed separately from corporate IT with restricted lateral movement
STEP 02

Locate target .fsa/.hid files

The attacker identifies the output directory where the Data Collection Software writes results. Default paths are well-documented in Thermo Fisher manuals. Files are named with run metadata (plate barcode, well position, date). An insider would know the directory structure; an external attacker would need to enumerate shares.
Conditions required:
  • Read access to the instrument output directory
  • Knowledge of the file naming convention or directory structure
Where this breaks in practice:
  • File naming is instrument-specific and not publicly standardized across all deployments
  • Some labs use LIMS middleware that moves files out of default locations
Detection/coverage: File-access audit logs (Windows Security Event 4663) on the output directory would record enumeration, but this auditing is rarely enabled on lab workstations by default.
STEP 03

Modify file contents

Using a hex editor, Python script, or the researcher-demonstrated tooling, the attacker opens the binary .fsa/.hid file and alters peak data, allele calls, or sample metadata. Nathan Adams (Forensic Bioinformatics) demonstrated combining two individuals' profiles into a single file that appeared untouched since 2015 — the operation took approximately 45 minutes with purpose-built tooling. No public exploit code has been released, but the file format is a well-documented ABIF (Applied Biosystems Inc. Format) binary structure.
Conditions required:
  • Write access to the target file
  • Understanding of ABIF binary format (publicly documented)
  • ~45 minutes of uninterrupted access per file
Where this breaks in practice:
  • Requires specialized domain knowledge of CE electropherogram data
  • Each file must be individually crafted — no mass-exploitation tooling exists
  • Chain-of-custody procedures in accredited labs require sign-off on file access
Detection/coverage: No detection mechanism exists in the unpatched software. The instrument and GeneMapper ID-X will load and process the tampered file without any warning or error. File integrity monitoring (FIM) tools like OSSEC, Wazuh, or Tripwire would detect the modification if deployed on the output directory, but this is uncommon in lab environments.
STEP 04

Tampered file enters analysis pipeline

The modified .fsa/.hid file is opened in GeneMapper ID-X or equivalent analysis software. Because the software performs no integrity verification, it treats the tampered data as authentic instrument output. The analyst sees fabricated allele calls and peak heights indistinguishable from legitimate data.
Conditions required:
  • Analyst opens the tampered file in the normal workflow
  • No out-of-band integrity check (hash comparison, FIM alert) catches the change
Where this breaks in practice:
  • Experienced analysts may notice biologically implausible results (e.g., three-allele loci in a single-source sample)
  • Peer review and technical review steps mandated by SWGDAM guidelines add a human detection layer
STEP 05

Falsified results influence outcomes

The tampered DNA profile is reported as a match, exclusion, or inconclusive result. In a forensic context this could exonerate the guilty or implicate the innocent. In a clinical context it could produce a misdiagnosis (e.g., incorrect paternity, wrong pharmacogenomic variant). The impact is bounded to the individual case but the consequence is irreversible once a report is issued.
Conditions required:
  • Tampered results survive peer review
  • Report is issued based on fabricated data
Where this breaks in practice:
  • Accredited labs require two-person review of all casework
  • Reanalysis of the original sample from retained biological material would reveal the discrepancy
  • Defense counsel increasingly requests raw data review in litigation
03 · Compensating Control

1
HIGH 7.5→MEDIUM 5.5
SEVERITY REDUCED
Deploy file integrity monitoring on all .fsa/.hid output directories — Configure FIM (Wazuh, OSSEC, Tripwire, or Windows SACL auditing with Event ID 4663) on every directory where Data Collection Software writes output files. Alert on any write operation not originating from the instrument software process. This is the single highest-value control because it directly detects the attack at Step 3. Deploy within 30 days per the noisgate mitigation SLA for HIGH-severity findings.
2
HIGH 7.5→LOW 3.0
SEVERITY REDUCED
Implement cryptographic hashing at file creation — Immediately after each instrument run, compute SHA-256 hashes of all generated .fsa/.hid files and store hashes in a separate, write-protected log (WORM storage, LIMS database, or blockchain-anchored notary). Compare hashes before any file is opened for analysis. This provides integrity assurance independent of the vendor's upcoming digital-signature feature. Deploy within 30 days.
3
HIGH 7.5→MEDIUM 5.0
SEVERITY REDUCED
Restrict write access to output directories — Apply NTFS ACLs so that only the Data Collection Software service account can write to output directories. All analyst accounts should have read-only access. This breaks the attack at Step 3 for any attacker who does not have the service account credential. Implement within 30 days.
4
HIGH 7.5→LOW 2.5
SEVERITY REDUCED
Apply vendor patches (digital signatures) — Update to 3500/3500xL 4.0.3, 3730/3730xL 5.0.3, SeqStudio 1.2.6, SeqStudio Flex 1.2.1, and GeneMapper ID-X 1.7.4. The patches add digital signatures to newly generated files, enabling the software to detect post-creation tampering. Note: patches do not protect historical files. For EOL platforms (3130, 3100, 310), plan instrument replacement. Remediate within 180 days per the noisgate remediation SLA for HIGH-severity findings.
5
HIGH 7.5→HIGH 6.5
Network-isolate instrument workstations — Ensure all genetic analyzer workstations and their file shares are on a dedicated VLAN with no routing to the corporate network. Allow only outbound connections required for LIMS integration via a jump host. This raises the bar for Step 1 significantly. Deploy within 30 days.
What doesn't work
  • Antivirus / EDR — this is not malware execution. The attacker is editing a data file with legitimate tools (hex editor, Python). No executable payload is delivered and no malicious process is spawned. EDR behavioral detection will not trigger.
  • Network IDS/IPS — the attack vector is local file modification, not network traffic. There is no network signature to detect.
  • Backup and restore — restoring from backup does not help if the backup contains the already-tampered file, or if the attacker modifies files between backup intervals. Backups are a recovery control, not a detection or prevention control for this vulnerability.
  • Standard disk encryption (BitLocker) — protects data at rest from offline theft but does not prevent a logged-in user or process from modifying files on a mounted volume.
04 · Intelligence Metadata

The supporting signals.

In-the-wild exploitationNone observed. Not listed in CISA KEV. No campaigns, advisories, or incident reports reference active exploitation of this vulnerability as of October 2026.
Proof of conceptPrivate demonstration only. Nathan Adams (Forensic Bioinformatics) demonstrated combining two DNA profiles into a single .fsa file that appeared unmodified since 2015, using purpose-built tooling in ~45 minutes. No public PoC code has been released on GitHub, ExploitDB, or pocindex.io. The ABIF file format is publicly documented, lowering the barrier for a motivated researcher.
EPSS0.00141 (0.141% probability of exploitation in the next 30 days). This places the CVE in the lower half of all scored vulnerabilities, consistent with local-access-only, niche-equipment flaws.
KEV statusNot listed. No CISA Known Exploited Vulnerabilities entry as of 2026-10-07.
CVSS vectorCVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H (8.4 HIGH). Note: the v3.1 vector overstates C and A impact — this is a pure integrity flaw. The CVSS 4.0 vector CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:N/VI:H/VA:N/SC:N/SI:H/SA:N (8.2 HIGH) is more accurate.
Affected versions3500/3500xL ≤4.0.2, 3730/3730xL ≤5.0.2, SeqStudio ≤1.2.5, SeqStudio Flex ≤1.2.0, GeneMapper ID-X ≤1.7.3, 3130 ≤4.1 (EOL), 3100/3100-Avant ≤2.0 (EOL), 310 ≤3.1 (EOL)
Fixed versions3500/3500xL 4.0.3, 3730/3730xL 5.0.3, SeqStudio 1.2.6, SeqStudio Flex 1.2.1, GeneMapper ID-X 1.7.4. Three legacy platforms (3130, 3100, 310) are end-of-life with no fix available.
Internet exposureNot applicable. These instruments do not expose network services to the internet. No Shodan, Censys, or GreyNoise hits for Applied Biosystems genetic analyzer interfaces. Attack surface is strictly local/LAN.
Disclosure date2026-08-05 via CISA ICS-CERT advisory ICSMA-26-216-01.
ResearchersNathan Adams, Kevin Dyer, and Laura Gaydosh Combs of Forensic Bioinformatics, coordinated through CISA.

Sources.

  1. CISA ICS-CERT Advisory ICSMA-26-216-01
  2. The Hacker News — Thermo Fisher Patches Flaw That Could Make DNA File Tampering Nearly Undetectable
  3. Hackread — Thermo Fisher Forensic DNA File Tampering Flaw
  4. Rescana — CVE-2026-17583 Analysis
  5. NVD — CVE-2026-17583
  6. Rapid7 — CVE-2026-17583
  7. Forensic Bioinformatics (researcher organization)
  8. Thermo Fisher — 3500 Series Genetic Analyzer for HID
05 · The Call

Final Verdict
= UNCHANGED to HIGH (7.5/10)

Why this verdict

  • Local access is a real gate. AV:L means the attacker must already be on the lab workstation or file share. Forensic labs operate on isolated networks per ISO 17025 and FBI QAS accreditation — lateral movement from corporate IT to the instrument subnet is a non-trivial prerequisite that implies a prior compromise stage.
  • Extremely narrow installed base. These are specialized capillary electrophoresis instruments deployed in forensic crime labs, paternity labs, and clinical genetics facilities — not general enterprise IT. The global population of affected systems is in the low thousands, not millions. This compresses the reachable target population dramatically.
  • No public exploit tooling. The researcher demonstrated capability but did not release code. The ABIF format is documented, but crafting a convincing tampered file requires domain expertise in forensic DNA analysis. This is not a spray-and-pray vulnerability.
  • EPSS confirms low exploitation probability. At 0.00141, the model places this well below the threshold where automated exploitation campaigns are expected. No KEV listing, no threat-intel signals.
  • Role multiplier: forensic/clinical DNA lab (OT-safety impact). Virtually 100% of Applied Biosystems Genetic Analyzer deployments occupy safety-critical roles — forensic evidence analysis for criminal proceedings, clinical diagnostics, and paternity testing. If the chain succeeds, the blast radius is *per-case* (individual evidence file), not fleet-scale, but the consequence is irreversible miscarriage of justice or clinical harm. This is a textbook OT-safety impact scenario. The verdict floor is HIGH because the affected component canonically occupies a high-value role and the safety impact is real, even though the blast radius does not reach domain/fleet/supply-chain scale.
  • Historical exposure is unmitigable. Files generated before the patch cannot be retroactively validated. The vulnerability has existed since ~1995. This long tail of unprovable integrity is a unique aggravating factor not captured by CVSS, though it does not change the *forward-looking* exploitability score.

Why not higher?

The attack requires local access to an isolated lab network, specialized knowledge of DNA electropherogram data, and per-file manual crafting with no mass-exploitation tooling. The blast radius is bounded to individual case files, not fleet or domain scale. EPSS is extremely low and there is zero evidence of in-the-wild exploitation or public exploit code. These friction points collectively prevent a CRITICAL rating despite the safety-critical deployment context.

Why not lower?

Nearly 100% of deployments are in forensic and clinical roles where successful exploitation produces irreversible real-world harm (wrongful conviction, exoneration of the guilty, clinical misdiagnosis). The OT-safety impact floor holds at HIGH. Three end-of-life platforms will never receive a fix, extending exposure indefinitely. The file format has been unsigned since 1995, creating a decades-long window of unprovable file integrity. Downgrading to MEDIUM would understate the consequence severity for this class of equipment.

06 · Verification

Crowdsourced verification payload.

Run this script on each genetic analyzer workstation with administrator privileges. It checks the Windows registry and common installation paths for affected Thermo Fisher Applied Biosystems software and compares installed versions against patched baselines. Invoke: powershell -ExecutionPolicy Bypass -File .\Check-CVE-2026-17583.ps1

noisgate-verify.ps1
POWERSHELLREAD-ONLYSAFE
#Requires -RunAsAdministrator
<#
  Check-CVE-2026-17583.ps1
  Detects whether Thermo Fisher Applied Biosystems Data Collection
  or GeneMapper ID-X software is installed at a vulnerable version.
  Exit codes: 1 = VULNERABLE, 0 = PATCHED, 2 = UNKNOWN
#>

$patchedVersions = @{
    '3500 Series Data Collection Software'        = [version]'4.0.3'
    '3500xL Series Data Collection Software'      = [version]'4.0.3'
    '3730 Series Data Collection Software'        = [version]'5.0.3'
    '3730xL Series Data Collection Software'      = [version]'5.0.3'
    'SeqStudio Genetic Analyzer'                  = [version]'1.2.6'
    'SeqStudio Flex Series Instrument Software'   = [version]'1.2.1'
    'GeneMapper ID-X'                             = [version]'1.7.4'
}

# EOL products with no fix available
$eolProducts = @(
    '3130 Series Data Collection Software',
    'ABI PRISM 3100',
    'ABI PRISM 3100-Avant',
    'ABI PRISM 310'
)

$found = $false
$vulnerable = $false

# Search Uninstall registry keys
$regPaths = @(
    'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall\*',
    'HKLM:\SOFTWARE\WOW6432Node\Microsoft\Windows\CurrentVersion\Uninstall\*'
)

foreach ($path in $regPaths) {
    Get-ItemProperty -Path $path -ErrorAction SilentlyContinue | ForEach-Object {
        $displayName = $_.DisplayName
        if (-not $displayName) { return }

        # Check EOL products first
        foreach ($eol in $eolProducts) {
            if ($displayName -match [regex]::Escape($eol)) {
                $found = $true
                $vulnerable = $true
                Write-Host "[!] EOL PRODUCT FOUND: $displayName (version $($_.DisplayVersion)) - NO PATCH AVAILABLE" -ForegroundColor Red
            }
        }

        # Check patchable products
        foreach ($product in $patchedVersions.Keys) {
            if ($displayName -match [regex]::Escape($product)) {
                $found = $true
                $installedStr = $_.DisplayVersion
                try {
                    $installed = [version]$installedStr
                    $required  = $patchedVersions[$product]
                    if ($installed -lt $required) {
                        $vulnerable = $true
                        Write-Host "[!] VULNERABLE: $displayName v$installed (need >= v$required)" -ForegroundColor Red
                    } else {
                        Write-Host "[+] PATCHED: $displayName v$installed (>= v$required)" -ForegroundColor Green
                    }
                } catch {
                    Write-Host "[?] FOUND but version unparseable: $displayName ($installedStr)" -ForegroundColor Yellow
                }
            }
        }
    }
}

# Also scan common install directories
$searchPaths = @(
    'C:\Applied Biosystems\*',
    'C:\Program Files\Applied Biosystems\*',
    'C:\Program Files (x86)\Applied Biosystems\*',
    'C:\Thermo Fisher Scientific\*'
)

foreach ($sp in $searchPaths) {
    if (Test-Path (Split-Path $sp -Parent)) {
        $dirs = Get-ChildItem -Path $sp -Directory -ErrorAction SilentlyContinue
        foreach ($d in $dirs) {
            if (-not $found) {
                Write-Host "[i] Applied Biosystems installation directory detected: $($d.FullName)" -ForegroundColor Cyan
                Write-Host "    Manual version check recommended if registry detection missed this product." -ForegroundColor Cyan
            }
        }
    }
}

if (-not $found) {
    Write-Host "UNKNOWN - No Thermo Fisher Applied Biosystems software detected on this host." -ForegroundColor Yellow
    exit 2
} elseif ($vulnerable) {
    Write-Host "`nVULNERABLE - One or more products require patching for CVE-2026-17583." -ForegroundColor Red
    exit 1
} else {
    Write-Host "`nPATCHED - All detected products are at or above the fixed version." -ForegroundColor Green
    exit 0
}