← Back to Feed CACHED · 2026-11-04 00:47:27 · CACHE_KEY CVE-2026-28336
CVE-2026-28336 · CWE-20 · Disclosed 2026-11-03

SolarWinds Dameware Remote Support authenticated command injection

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

A helpdesk tech with a lowly account gets the keys to every machine they can click on

SolarWinds Dameware Remote Support / Mini Remote Control contains a command injection flaw (CWE-20). An authenticated low-privilege user who can reach the Dameware console can pass shell metacharacters through a parameter used in a remote session or administrative action; the unsanitized value is executed by the OS on the remotely controlled host, giving the attacker SYSTEM- or root-level code execution. Affected versions are 2026.3 and earlier; fixed in 2026.4.

SolarWinds rates this 9.9 CRITICAL with AV:N/PR:L/UI:N/S:C. That score presumes the Dameware console is network-reachable by a low-privileged account and that compromise of one remote host is a scope change. In most enterprise deployments the console sits on an internal management VLAN and requires valid low-privilege credentials — a real barrier. No public PoC, Metasploit/Exploit-DB/nuclei template, or in-the-wild exploitation has been observed as of 2026-11-03, so the vendor’s CRITICAL is too high for today’s exposure, but the post-auth blast radius is still severe.

Bottom line: Patch to 2026.4 on the noisgate remediation SLA (≤180 days for HIGH), and immediately restrict Dameware console/agent network paths and enforce MFA/least privilege on the noisgate mitigation SLA (≤30 days). Treat any low-priv Dameware account as a potential SYSTEM pivot.

"An authenticated low-priv user can turn a Dameware session into a SYSTEM shell on the remote host — but lack of public exploit and internal exposure keep this out of CRITICAL today."
02 · The Attack Path

5 steps from start to impact.

STEP 01

Gain low-privilege access to the Dameware console

Attacker obtains credentials for a low-privilege Dameware user — via phishing, credential stuffing, or existing insider access. The Dameware Remote Support central console is the weaponized tool here; no exploit code is needed for this step. They connect over TCP to the internal Dameware service.
Conditions required:
  • Valid low-privilege account in SolarWinds Dameware
  • Network access to the Dameware server (usually internal/VPN)
Where this breaks in practice:
  • Dameware consoles are rarely published to the internet
  • Low-privilege Dameware accounts may be tied to AD and can require MFA
  • No known default credentials for current versions
Detection/coverage: Dameware login audit logs; SIEM alerts on impossible-travel or failed login storms.
STEP 02

Locate a remote host with a Dameware agent and initiate an action

The attacker browses the managed endpoint list in the console and selects a target host that has the Dameware agent installed. The Dameware Mini Remote Control / Remote Support agent is the execution vehicle. They initiate a session or administrative task that calls an external process with parameters.
Conditions required:
  • Target host has Dameware agent or is reachable via Dameware proxy
  • Attacker has rights to initiate remote-control actions
Where this breaks in practice:
  • Role-based restrictions may block certain actions for low-priv users
  • Agents are not installed on all hosts; many security teams exclude domain controllers
Detection/coverage: Dameware session logs; unexpected remote session to critical server alerts.
STEP 03

Inject shell metacharacters into an unsanitized parameter

In a field such as hostname, path, or pre-session command, the attacker appends shell metacharacters — e.g., ; calc, | powershell -e ... — and submits the request. Because CVE-2026-28336 exists in versions ≤2026.3, the value is passed to a shell or command interpreter on the remote host without escaping, executing the supplied command as the agent service account.
Conditions required:
  • Dameware version ≤2026.3
  • Specific vulnerable action/parameter is reachable by the authenticated user
  • Remote host OS supports shell metacharacters
Where this breaks in practice:
  • No public PoC/exploit module exists as of 2026-11-03, so the exact parameter must be found manually or via private research
  • Input validation differences may vary by deployment
Detection/coverage: EDR/Sysmon on remote host can flag unexpected child processes from Dameware agent (e.g., cmd.exe, powershell.exe).
STEP 04

Gain SYSTEM/root on the remote host

The Dameware agent typically runs as a high-privilege local service, so injected commands execute as SYSTEM on Windows or root on Unix-managed hosts. The attacker establishes persistence or exfiltrates data, and the remotely controlled host is fully compromised.
Conditions required:
  • Agent service runs with elevated privileges (standard default)
  • No application whitelisting blocks the child process
Where this breaks in practice:
  • On hardened endpoints, AppLocker/WDAC or EDR may kill the shell
  • Some deployments run the agent as a lower-priv account (less common)
Detection/coverage: EDR process tree, Sysmon Event ID 1, endpoint detection rules for Dameware spawning shells.
STEP 05

Pivot to domain or fleet impact

From a compromised high-value remote host — a domain controller, file server, or SQL server — the attacker dumps credentials, moves laterally, or deploys ransomware. This is where the RMM-like deployment role multiplier turns a single host compromise into domain or fleet-scale impact.
Conditions required:
  • Target host is high-value (DC, server) and has Dameware agent installed
  • Weak network segmentation allows lateral traffic
  • No tiered admin / privileged access management for domain trusts
Where this breaks in practice:
  • Many orgs exclude DCs from remote-control tooling as a security baseline
  • Credential dumping triggers EDR/identity protection alerts if configured
Detection/coverage: SIEM correlation, Defender for Identity/AD audit, lateral movement detections.
03 · Compensating Control

1
HIGH 8.0→IGNORE 0.0
SEVERITY REDUCED
Upgrade SolarWinds Dameware to 2026.4 or later — Apply the vendor patch on the Dameware central server, proxy, and all agents/remote clients. This removes the command injection sink; because it is the full remediation, aim to complete it within the noisgate remediation SLA for HIGH: ≤180 days.
2
HIGH 8.0→HIGH 7.5
Restrict Dameware console and agent traffic to a dedicated admin VLAN and jump hosts — Put the Dameware server behind a management network and restrict access to named admin workstations via firewall policy. This breaks the AV:N assumption for external or broad internal access. Deploy within 30 days on the noisgate mitigation SLA for HIGH.
3
HIGH 8.0→HIGH 7.0
Enforce MFA and least privilege on all Dameware accounts, and audit low-priv users — Require MFA for Dameware console logins and remove/disable unnecessary low-priv accounts. This reduces the pool of attackers who can reach the vulnerable function. Implement within 30 days.
4
HIGH 8.0→MEDIUM 5.5
SEVERITY REDUCED
Deploy EDR/Sysmon on managed hosts to detect and block suspicious child processes from Dameware agents — Configure EDR to kill/alert on cmd.exe, powershell.exe, cscript.exe, or other shells spawned by the Dameware agent. This doesn't patch the injection but blunts SYSTEM execution. Implement within 30 days.
What doesn't work
  • Network microsegmentation alone: already-low-priv console users are still inside the trusted zone, so this doesn't stop the post-auth injection.
  • Standard WAF/IDPS on the perimeter: Dameware traffic is normally internal and uses proprietary protocols; external web filters won't see the injection.
  • Unauthenticated vulnerability scanning: cannot confirm this flaw because the attack requires valid credentials and a non-HTTP session.
04 · Intelligence Metadata

The supporting signals.

In-the-wild statusNo known exploitation; OffSeq Mirage saw no traffic; not in CISA KEV; no vendor or third-party reports of active use as of 2026-11-03.
Proof-of-concept availabilityNo public exploit code indexed by Metasploit, Exploit-DB, GitHub, or nuclei templates; Mallory/OffSeq report no PoC.
EPSS score / percentileNo EPSS yet (Feedly: No EPSS yet); treat as low probability but not zero.
KEV status / datesNot listed in CISA KEV; no known exploitation windows.
CVSS vectorCVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H — 9.9 CRITICAL; scope change reflects pivot from console to remote host.
Affected versionsDameware 2026.3 and all earlier versions.
Fixed versionsSolarWinds Dameware 2026.4 or later.
Scanning / exposure dataNo public Shodan/Censys/GreyNoise data; Dameware consoles are typically internal-only, but internet-facing instances should be assumed rare.
Disclosed / reportedCVE published 2026-11-03; reporting researcher/org not disclosed in available sources.

Sources.

  1. VulDB — CVE-2026-28336
  2. CVE.org Record — CVE-2026-28336
  3. Rapid7 Vulnerability Database — CVE-2026-28336
  4. Mallory — Authenticated Command Injection in SolarWinds Dameware
  5. OffSeq Threat Radar — CVE-2026-28336
  6. Feedly — CVE-2026-28336
05 · The Call

Final Verdict
↓ DOWNGRADED to HIGH (8.0/10)

Why this verdict

  • Role multiplier: Dameware is an RMM-style reach-into-everything platform. In the low-value role, a compromised workstation is just a workstation; in the typical role, it manages line-of-business servers; in the high-value role, Dameware agents are installed on domain controllers, file servers, or the IT management plane itself. In that high-value role the post-auth chain gives a low-priv console user SYSTEM on a DC → DA → domain takeover. We estimate ≥10% of enterprise Dameware deployments include at least one server/DC agent, so the floor is HIGH.
  • Authentication and network reach: CVSS treats this as AV:N, but the first hop requires a valid low-priv account and reachability to the Dameware console, which is normally hidden behind a VPN/internal VLAN. This removes the internet-drive-by mass-exploitation scenario and pulls the score down from 9.9.
  • No public exploit or active use: As of 2026-11-03 there is no PoC, no Metasploit/Exploit-DB/nuclei template, and no KEV or in-the-wild reports, so the likelihood of immediate automated exploitation is lower than the vendor score suggests.
  • Patch available: A fixed release (2026.4) exists, so this is a solvable defect, not an unpatched zero-day.

Why not higher?

Not CRITICAL today because the documented chain requires valid low-privilege credentials and access to an internal management console; there is no unauthenticated remote vector, no public exploit, and no KEV/exploited-in-the-wild signal. The 9.9 CVSS score assumes network reach and immediate scope change, which overstates the number of exposed orgs.

Why not lower?

Not MEDIUM because the post-auth payoff is SYSTEM on the remote host and, in the high-value role, can lead to domain or fleet compromise. The low-privilege bar is not high enough to offset that multiplier; also, remote support agents are often installed broadly across servers.

06 · Verification

Crowdsourced verification payload.

Run on any machine where SolarWinds Dameware Remote Support, Mini Remote Control, or Internet Session Agent is installed — central servers, admin workstations, and managed endpoints. Requires local administrator rights on the target host. Invoke from PowerShell as powershell -ExecutionPolicy Bypass -File check_dameware.ps1.

noisgate-verify.ps1
POWERSHELLREAD-ONLYSAFE
# check_dameware.ps1
# Checks installed SolarWinds Dameware versions against CVE-2026-28336
# Output: VULNERABLE (version <= 2026.3), PATCHED (>= 2026.4), UNKNOWN
# Exit: 1=VULNERABLE, 0=PATCHED, 2=UNKNOWN
$ErrorActionPreference = 'SilentlyContinue'
$paths = @(
  'HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall\*',
  'HKLM:\SOFTWARE\WOW6432Node\Microsoft\Windows\CurrentVersion\Uninstall\*'
)
$entries = foreach ($p in $paths) {
  Get-ItemProperty -Path $p -ErrorAction SilentlyContinue |
    Where-Object { $_.DisplayName -match 'Dameware' }
}
if (-not $entries) {
  Write-Output 'UNKNOWN'
  exit 2
}
$vulnerableFound = $false
$patchedFound = $false
foreach ($e in $entries) {
  $ver = $e.DisplayVersion
  if ($ver -and $ver -match '^(\d{4})(?:\.(\d+))?') {
    $year = [int]$Matches[1]
    $minor = [int]$Matches[2]
    if ($year -lt 2026 -or ($year -eq 2026 -and $minor -le 3)) {
      $vulnerableFound = $true
    }
    elseif ($year -gt 2026 -or ($year -eq 2026 -and $minor -ge 4)) {
      $patchedFound = $true
    }
  }
}
if ($vulnerableFound) { Write-Output 'VULNERABLE'; exit 1 }
elseif ($patchedFound) { Write-Output 'PATCHED'; exit 0 }
else { Write-Output 'UNKNOWN'; exit 2 }