Introduction
FortiManager and FortiAnalyzer sit at the center of many Fortinet deployments, managing fleets of security devices and aggregating their telemetry, so a flaw in how these appliances build signed SAML responses touches the trust model of everything they oversee. Fortinet's advisory FG-IR-26-222 describes exactly that: an unauthenticated remote attacker may be able to inject arbitrary content into a SAML response that the appliance then signs.
The issue is tracked as CVE-2026-126176 and carries a CVSS 3.1 base score of 8.1 (High), assigned by Fortinet as the CNA. NVD identifies the vendor as Fortinet, Inc. (the original vendor field was listed as Unknown) and has not published its own independent score. Fixed firmware is available for all four affected product families: FortiAnalyzer, FortiAnalyzer Cloud, FortiManager and FortiManager Cloud.
Technical information
Root cause
Fortinet classifies the weakness as CWE-91, XML Injection (aka Blind XPath Injection). The advisory names the affected component as API and the attack type as Unauthenticated. Its core technical statement is that the flaw "may allow an unauthenticated remote attacker to inject arbitary content into a signed SAML response via crafted SAML requests."
MITRE defines CWE-91 as improper neutralization of special XML elements, which lets an attacker modify XML syntax, content or commands before another system processes them. Applied to this advisory, the wording indicates that the vulnerable code path was building a SAML response, copying attacker-supplied data from the incoming SAML request into that XML without neutralizing it, and then signing the result.
Why the signature matters
The injected content lands in a signed SAML response, and that matters more than the malformed XML itself. A valid signature tells downstream consumers to trust the whole signed structure. Anything injected before signing is covered by that signature, so it can alter the meaning of the response while still appearing legitimate to a verifier.
This is also why the issue is different from classic signature wrapping: nothing in the advisory suggests the signature is broken or bypassed. The appliance itself signs content it did not intend to produce.
Attack flow, as far as the advisory supports it
The advisory supports a high-level sequence, not a reconstructed exploit chain:
- An unauthenticated attacker with network access to an affected appliance sends a crafted SAML request.
- Content in that request is not neutralized and is incorporated into the SAML response XML the appliance generates.
- The appliance signs the response, including the injected content.
- The resulting signed SAML response carries attacker-controlled content backed by a valid signature.
Fortinet does not disclose which request field is affected, which response element changes, the endpoint path or port involved, whether the appliance is acting in an identity provider or service provider role, or any configuration prerequisites. The available material also does not establish a demonstrated administrator login bypass, credential exposure or remote code execution. We have kept the attack flow at that level deliberately.
CVSS interpretation
The published base vector is:
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:L/I:H/A:L
- Attack vector: Network
- Attack complexity: High
- Privileges required: None
- User interaction: None
- Scope: Changed
- Confidentiality / Integrity / Availability: Low / High / Low
Fortinet also publishes a 7.7 temporal score with E:P/RL:O/RC:C. Under the FIRST CVSS 3.1 specification, E:P is the Proof of Concept maturity category, RL:O means an official fix exists, and RC:C indicates confirmed reporting confidence.
The High attack complexity rating signals that some condition beyond the attacker's control must be met, but the advisory does not say what it is. Fortinet's advisory also labels the impact as denial of service, while the vector rates integrity as High. Both facts stand, but neither supplies the missing exploit details, and neither justifies escalating the finding to confirmed account takeover.
Exploitation status
Fortinet marks the advisory Known Exploited: No. A Health ISAC alert from November 11, 2026 reports no active exploitation of the broader batch of Fortinet issues it covers. That alert's title references arbitrary code execution, but within it CVE-2026-126176 is listed specifically as XML Injection, so the batch's worst-case impact should not be assigned to this CVE. VulDB lists public exploit availability as no. Despite the E:P temporal metric, we did not identify a public exploit, a named threat actor, or CVE-specific indicators of compromise.
Patch information
Vendor firmware, not a public code diff
Fortinet fixed CVE-2026-126176 in new firmware builds for FortiAnalyzer, FortiAnalyzer Cloud, FortiManager and FortiManager Cloud, documented in PSIRT advisory FG-IR-26-222 (published 2026-11-10). NVD lists that advisory as its only reference for this CVE. These are closed source appliances, so there is no public commit, diff or source snippet to review. The fixed release numbers are the only authoritative information about the patch.
Fixed versions by branch
- FortiAnalyzer / FortiManager 8.0: 8.0.0 vulnerable, fixed in 8.0.1 or above
- FortiAnalyzer / FortiManager 7.6: 7.6.0 through 7.6.7 vulnerable, fixed in 7.6.8 or above
- FortiAnalyzer / FortiManager 7.4: 7.4.0 through 7.4.11 vulnerable, fixed in 7.4.12 or above
- FortiAnalyzer / FortiManager 7.2: 7.2.0 through 7.2.12 vulnerable, fixed in 7.2.13 or above
- FortiAnalyzer Cloud / FortiManager Cloud 8.0: 8.0.0 vulnerable, fixed in 8.0.1 or above
- FortiAnalyzer Cloud / FortiManager Cloud 7.6: 7.6.0 through 7.6.7 vulnerable, fixed in 7.6.8 or above
- FortiAnalyzer Cloud / FortiManager Cloud 7.4: 7.4.1 through 7.4.11 vulnerable, fixed in 7.4.12 or above
- FortiAnalyzer Cloud / FortiManager Cloud 7.2: 7.2.1 through 7.2.12 vulnerable, fixed in 7.2.13 or above
Each supported branch has its own fixed build, and you need at least the fixed build on the branch you are running. A 7.4.11 system needs 7.4.12; moving to 7.6.7 would not help.
Note on the 8.0 branch: FG-IR-26-222 lists 8.0.1 as the fixed build for this CVE, while Fortinet's October hardening release table lists 8.0.2 for that release batch as a whole. Moving to 8.0.2 or later covers both.
What the fix has to change
Fortinet has not published the code-level change. For this class of bug (CWE-91 in a SAML signing path), a fix would normally:
- escape or strictly validate every request-derived value (for example, IDs and URLs echoed back from the AuthnRequest) before it goes into the response;
- build the XML with a DOM or serializer API instead of string concatenation;
- make sure the signed content is exactly the structure the appliance meant to produce.
That list is general guidance for CWE-91 in SAML signing, not something Fortinet has confirmed.
No virtual patch
The advisory lists Virtual Patch: No, so there is no FortiGuard IPS signature that blocks this flaw while you wait to upgrade. Installing the fixed firmware is the only vendor-provided fix. Fortinet's broader hardening guidance (restricting administrative HTTPS and SSH to trusted addresses, keeping FortiManager administration off the internet, strong credentials with multifactor authentication, and log monitoring) can reduce exposure, but Fortinet states these measures do not replace timely upgrades, and credential controls do not address an unauthenticated attack path.
Release timing
Fortinet credits the discovery to its own Product Security team, working with a frontier AI model. The fix shipped through Fortinet's security hardening release program, which delivered FortiManager and FortiAnalyzer 7.6.8, 7.4.12 and 7.2.13 on October 22, 2026, about three weeks before the November 10 advisory. Those are the same fixed builds FG-IR-26-222 lists for the 7.x branches. Fortinet's blog notes that released updates can show attackers where a weakness is, so administrators should upgrade before attackers can compare old and new firmware to locate the bug.
Detection methods
Right now, CVE-2026-126176 can only be detected by checking versions. Scanners can tell you which appliances run vulnerable builds, but no public signatures exist to show whether one of them has been attacked. FG-IR-26-222 contains no IoCs and lists Virtual Patch: No, meaning no FortiGuard IPS signature was released with it, so a FortiGate in front of these appliances cannot be relied on to flag exploit traffic. We found no CVE-specific Snort, Suricata, Sigma, YARA or Nuclei content.
Tenable Nessus
Tenable's pipeline entry for FG-IR-26-222 (issue 236608) has status Released and lists two plugins, both published 11/11/2026, each covering CVE-2026-126176 and its sibling CVE-2026-126177:
- Plugin 378654, Fortinet FortiAnalyzer XML injection (FG-IR-26-222), file
fortianalyzer_FG-IR-26-222.nasl - Plugin 378655, Fortinet FortiManager XML injection (FG-IR-26-222), file
fortimanager_FG-IR-26-222.nasl
Both are Local checks in the Firewalls family. They require the KB items Host/Fortigate/model and Host/Fortigate/version, so the scan must log in to the appliance to collect model and version. An unauthenticated network sweep will not populate those items, and the plugins will not fire. Tenable notes that Nessus does not test the flaw itself and relies only on the self-reported version.
Two consequences follow:
- A finding means the version is in the affected range, not that the appliance was exploited.
- Missing credentials can hide exposure. A clean result on a host the scanner could not log in to tells you nothing.
Rapid7 InsightVM / Nexpose
Rapid7's database lists CVE-2026-126176 as covered for Fortinet FortiManager, added Nov 11, 2026, with solutions mapping to 7.2.13, 7.4.12, 7.6.8 and 8.0.1. The entry names only FortiManager, so do not assume Rapid7 flags FortiAnalyzer. Use Nessus or a manual version check for those appliances.
Manual inventory check
Compare each appliance's firmware against the FG-IR-26-222 affected ranges:
- FortiAnalyzer and FortiManager: 8.0.0, 7.6.0 through 7.6.7, 7.4.0 through 7.4.11, and 7.2.0 through 7.2.12
- Cloud editions: the same, except the 7.4 branch starts at 7.4.1 and the 7.2 branch at 7.2.1
Anything in those ranges is exposed.
Threat hunting focus
No vendor-published pattern exists yet, but the advisory description indicates where to look:
- Prioritize appliances where SAML SSO is enabled and the login interface is reachable from untrusted networks.
- Review SAML-authenticated admin logins. Pay closest attention to logins without a corresponding session at your identity provider, logins from unexpected source addresses, and logins as accounts that rarely use the GUI.
- Because the vector rates integrity impact High, look for unexplained admin, configuration or device policy changes following such logins.
This is hunting guidance derived from the vulnerability description, not a published signature. Test it against your own normal SAML traffic before turning it into alerts.
Affected systems and versions
Per FG-IR-26-222 and the NVD record:
FortiAnalyzer
- 8.0.0
- 7.6.0 through 7.6.7
- 7.4.0 through 7.4.11
- 7.2.0 through 7.2.12
FortiAnalyzer Cloud
- 8.0.0
- 7.6.0 through 7.6.7
- 7.4.1 through 7.4.11
- 7.2.1 through 7.2.12
FortiManager
- 8.0.0
- 7.6.0 through 7.6.7
- 7.4.0 through 7.4.11
- 7.2.0 through 7.2.12
FortiManager Cloud
- 8.0.0
- 7.6.0 through 7.6.7
- 7.4.1 through 7.4.11
- 7.2.1 through 7.2.12
The fixed targets are identical across appliance and Cloud editions, but the lower bounds differ on the 7.4 and 7.2 branches. Cloud editions are explicitly included, and the advisory provides no automatic remediation assurance for them, so they belong in your inventory alongside appliances. The advisory does not disclose specific vulnerable configurations. The detection guidance above treats SAML SSO-enabled appliances reachable from untrusted networks as the highest priority for review.
Vendor security history
Fortinet states it serves more than 900,000 customers with over 50 enterprise-grade products. FortiManager centrally manages Fortinet network security devices, and FortiAnalyzer aggregates logs and telemetry from Fortinet and third-party systems for analytics and correlation across the Security Fabric.
Management platforms from Fortinet have been targeted before. Mandiant reported that UNC5820 exploited the separate FortiManager vulnerability CVE-2024-47575 as early as June 27, 2024, and its investigation with Fortinet involved 50+ potentially compromised FortiManager devices across industries. In that incident, the actor staged and exfiltrated configuration data from managed FortiGate devices, including users and hashed passwords. Mandiant did not find evidence in its analyzed data that the information was used for further compromise.
That case shows why these platforms warrant prompt patching, but it is a different vulnerability and a different evidence state. Nothing in the reviewed material links UNC5820 or its indicators to CVE-2026-126176. For the current issue, Fortinet's response combined internal discovery, fixed builds shipped ahead of public disclosure, and interim exposure guidance.
References
- NVD: CVE-2026-126176
- CVE Record: CVE-2026-126176
- FortiGuard PSIRT Advisory FG-IR-26-222
- Fortinet: Security Hardening Release for Multiple Products
- Tenable Nessus Plugin 378654 (FortiAnalyzer)
- Tenable Nessus Plugin 378655 (FortiManager)
- Tenable Plugin Pipeline Issue 236608
- Rapid7 Vulnerability Database: CVE-2026-126176
- HKCERT: Fortinet Products Multiple Vulnerabilities
- Health ISAC Alert: Multiple Vulnerabilities in Fortinet Products
- VulDB: CVE-2026-126176 in FortiAnalyzer
- Mandiant: Investigating FortiManager Zero Day Exploitation (CVE-2024-47575)
- MITRE CWE-91: XML Injection (aka Blind XPath Injection)
- FIRST CVSS v3.1 Specification Document
- Fortinet: About Us
- FortiManager Data Sheet
- FortiAnalyzer Data Sheet
- CISA: Fortinet Releases Advisory on New Post Exploitation Technique for Known Vulnerabilities



