Automated Threat Modeling

IBM HTTP Server CVE-2026-8855: Code Injection via TLS Mutual Authentication — Brief Summary and Patch Guidance

A brief summary of CVE-2026-8855, a high-severity code injection vulnerability in IBM HTTP Server that affects TLS mutual authentication configurations. Includes affected versions, patch information, and remediation guidance.

CVE Analysis

10 min read

ZeroPath CVE Analysis
ZeroPath CVE Analysis

2026-05-26

IBM HTTP Server CVE-2026-8855: Code Injection via TLS Mutual Authentication — Brief Summary and Patch Guidance
Experimental AI-Generated Content

This CVE analysis is an experimental publication that is completely AI-generated. The content may contain errors or inaccuracies and is subject to change as more information becomes available. We are continuously refining our process.

If you have feedback, questions, or notice any errors, please reach out to us.

[email protected]

Introduction

A code injection vulnerability in IBM HTTP Server's TLS mutual authentication handling gives remote, unauthenticated attackers a path to both arbitrary code execution and denial of service on enterprise web infrastructure. Ironically, the flaw only affects deployments that have enabled client certificate authentication, a configuration found most often in the highest-security environments running under zero-trust principles.

CVE-2026-8855 was disclosed on May 26, 2026, in a bulletin covering nine vulnerabilities in IBM HTTP Server, including CVE-2026-9170 at a critical CVSS 9.8. IBM rates this flaw at 8.1 HIGH, while NIST independently scored it at 9.8 CRITICAL. There are no workarounds. The only remediation is patching.

Technical Information

Root Cause: CWE-94 Code Injection in Certificate Processing

CVE-2026-8855 is classified under CWE-94: Improper Control of Generation of Code ('Code Injection'). The flaw is in how IBM HTTP Server processes client certificate data during TLS mutual authentication. It is triggered when the SSLClientAuth directive is set to any value other than "0" or "None", which means the server requires or accepts client certificates as part of the TLS handshake.

In a mutual TLS deployment, the server has to parse values from the client certificate (such as the Subject Distinguished Name, Subject Alternative Names, and other X.509 fields) to make access-control decisions. CWE-94 describes a product that builds all or part of a code segment from externally influenced input but does not properly neutralize special elements that could change the syntax or behavior of that code. When IBM HTTP Server dynamically builds code segments from this certificate data without strict neutralization, it creates a direct injection vector.

CVSS Vector Breakdown

The CVSS v3.1 vector is CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H:

CVSS ComponentValueInterpretation
Attack Vector (AV)Network (N)Exploitable remotely over a network
Attack Complexity (AC)High (H)Requires TLS mutual auth to be configured
Privileges Required (PR)None (N)No authentication needed to exploit
User Interaction (UI)None (N)No user action required
Scope (S)Unchanged (U)Impact confined to the vulnerable component
Confidentiality (C)High (H)Complete information disclosure possible
Integrity (I)High (H)Complete system compromise possible
Availability (A)High (H)Complete denial of service possible

The High attack complexity reflects the prerequisite that TLS mutual authentication must be configured. IBM's interim fix documentation calls the SSLClientAuth mutual auth configuration "uncommon" in typical deployments, which explains the AC:H rating. For environments that do use client certificate authentication, though, the exposure is direct, and the rest of the vector is severe: no privileges, no user interaction, and High ratings for all three impacts.

The gap between IBM's 8.1 HIGH and NIST's independent 9.8 CRITICAL likely comes from a different reading of attack complexity. If NIST rated AC as Low, on the reasoning that an attacker only needs to connect to a server that already has mutual TLS enabled, the score would reach 9.8.

Attack Flow

Based on the CWE-94 classification and the TLS mutual authentication attack surface, the exploitation chain likely runs as follows:

  1. Reconnaissance: The attacker finds an IBM HTTP Server instance with client authentication enabled. This shows up during the TLS handshake, when the server sends a CertificateRequest message.

  2. Certificate Crafting: The attacker builds a malicious client certificate with specially crafted elements in its fields, designed to break out of the data context and into a code generation context inside the server.

  3. TLS Handshake: The attacker opens a TLS connection and presents the crafted certificate. The server processes the certificate fields as part of its authentication and access control logic.

  4. Code Injection: The server places the untrusted certificate field values into a dynamically generated code segment without adequately neutralizing special elements. This is the core CWE-94 weakness, in which control-plane data gets injected into what should be the data plane.

  5. Execution: The injected code runs in the server process context. It can lead to arbitrary command execution (the RCE path) or a crash (the DoS path).

The related CAPEC attack patterns include CAPEC-242 (Code Injection) and CAPEC-35 (Leverage Executable Code in Non Executable Files). CWE-94's documented consequences say injected code can "control authentication" and "bypass protection mechanisms," so an attacker can "gain privileges" and reach resources they are "directly prevented from accessing." A successful exploit here completely undermines the trust guarantees the TLS mutual authentication handshake was designed to provide.

Even if an attacker cannot get reliable, deterministic code execution, they may still trigger a denial of service, so the dual RCE and DoS impact leaves them a fallback.

Broader Bulletin Context

CVE-2026-8855 is one of nine vulnerabilities addressed in the same IBM Security Bulletin, which suggests systemic issues in the component:

CVE IDCVSS ScoreDescription
CVE-2026-91709.8 (Critical)Code injection (CWE-94) in IBM Web Server Plug-ins
CVE-2026-88558.1 (High)RCE and DoS in TLS mutual auth (CWE-94)
CVE-2026-8854HighIBM HTTP Server vulnerability
CVE-2026-8856HighIBM HTTP Server vulnerability
CVE-2026-8852HighIBM HTTP Server vulnerability
CVE-2026-8850HighIBM HTTP Server vulnerability
CVE-2026-8835HighIBM HTTP Server vulnerability
CVE-2026-8834HighIBM HTTP Server vulnerability
CVE-2026-45186HighIBM HTTP Server vulnerability

CVE-2026-9170 sits in the same bulletin with the same CWE-94 classification and a 9.8 CRITICAL score. That strongly suggests a systemic code injection problem across IBM HTTP Server and its plug-in components.

Patch Information

IBM published Security Bulletin 7274065 on May 26, 2026, confirming an official fix. There are no workarounds or mitigations. Applying the patch is the only remediation.

Fix Details

The fix is tracked under APAR PH71265 and delivered through the Cumulative Security Interim Fix (CSIF) IFPH71265. This interim fix supersedes all prior IHS security interim fixes (PH71061, PH70572, PH68462, and others). IBM services IHS on a cumulative model, where each new CSIF rolls up all previous security fixes, so administrators only need to apply the latest interim fix.

Remediation Paths

IBM HTTP Server 9.0 (versions 9.0.0.0 through 9.0.5.28):

  • Apply interim fix IFPH71265 on top of a supported fix pack level (9.0.5.25, 9.0.5.26, or 9.0.5.27)
  • Or wait for Fix Pack 9.0.5.29 (targeted availability 3Q2026)

IBM HTTP Server 8.5 (versions 8.5.0.0 through 8.5.5.29):

  • Apply interim fix IFPH71265 on top of a supported fix pack level (8.5.5.28 or 8.5.5.29)
  • Or wait for Fix Pack 8.5.5.30 (targeted availability 3Q2026)

Available Packages

IBM released the interim fix on May 26, 2026, through IBM Fix Central. The packages include:

PackagePlatform(s)
9.0.5.27-WS-WASIHS-IFPH71265All (via IM)
9.0.5.26-WS-WASIHS-IFPH71265All (via IM)
9.0.5.25-WS-WASIHS-IFPH71265All (via IM)
8.5.5.29-WS-WASIHS-IFPH71265All (via IM)
8.5.5.28-WS-WASIHS-IFPH71265All (via IM)
9.0.5-WS-IHS-ARCHIVE-*-FP027-IFPH71265Windows x86_64, Linux x86_64, Linux s390x, Linux ppc64le, AIX ppc64

Packages are available through IBM Installation Manager, and as standalone archive packages for deployments without IM.

Compensating Controls

IBM has stated that no workarounds exist, but organizations facing patching delays may consider these compensating measures (not endorsed by IBM):

  • Disable TLS mutual authentication where operationally feasible by setting SSLClientAuth to "0" or "None", which removes the exploitation prerequisite entirely. This may not be acceptable in zero-trust or regulated environments.
  • Network segmentation to restrict which clients can open TLS connections to the IBM HTTP Server.
  • WAF or IDS/IPS rules to detect anomalous client certificate fields, though their effectiveness against this specific exploit is unconfirmed.
  • Enhanced monitoring of IBM HTTP Server processes for unexpected code execution patterns or crash events.

Because the same bulletin includes CVE-2026-9170 at CVSS 9.8, organizations should treat all nine vulnerabilities as one bundled patching event.

Affected Systems and Versions

IBM HTTP Server

ProductAffected Version RangeFixed VersionPlatform
IBM HTTP Server V9.09.0.0.0 through 9.0.5.289.0.5.29 (or interim fix IFPH71265)AIX, Linux, Windows, z/OS
IBM HTTP Server V8.58.5.0.0 through 8.5.5.298.5.5.30 (or interim fix IFPH71265)AIX, Linux, Windows, z/OS

Bundled Products

IBM HTTP Server ships as a component of several IBM products, so the blast radius extends beyond standalone deployments:

ProductAffected Version Range
IBM WebSphere Application Server (all editions)Bundles affected HTTP Server versions
IBM DevOps Code ClearCase11.0.0 through 11.0.0.05
IBM Rational ClearCase10.0.0 through 10.0.1.05; 9.1.0 through 9.1.0.10

Vulnerable Configuration

The vulnerability is only exploitable when TLS mutual authentication (client authentication) is enabled, meaning the SSLClientAuth directive is set to any value other than "0" or "None". IBM calls this configuration "uncommon" in typical deployments, but it is standard practice in enterprises running zero-trust architectures, particularly in financial services, government, and healthcare.

Vendor Security History

IBM's 2026 security response for IBM HTTP Server shows a pattern of bundled multi-CVE disclosures. In May 2026 alone, IBM PSIRT issued multiple security bulletins:

  • A May 13, 2026 bulletin covering multiple vulnerabilities in IBM HTTP Server used by WebSphere Application Server that affect IBM Business Automation Workflow.
  • A May 18, 2026 bulletin addressing multiple vulnerabilities in IBM HTTP Server shipped with IBM DevOps Code ClearCase.
  • Bulletins addressing IBM WebSphere Liberty vulnerabilities, including SSRF (CVE-2026-1561) and hard-coded cryptographic key issues.

The current bulletin addresses nine CVEs at once, with one critical and multiple high-severity issues sharing the same CWE-94 classification, which suggests accumulated security debt in the HTTP Server component. No workarounds are available for any of the nine CVEs, which indicates the vulnerabilities are deeply embedded in the server's code and cannot be mitigated through configuration changes alone.

The Hong Kong CERT (HKCERT) also issued an advisory on May 20, 2026, covering multiple vulnerabilities in IBM WebSphere products, another sign of how broad the issue is.

References

Detect & fix
what others miss

Works with
  • GitHub
  • GitLab
  • Bitbucket
  • Azure DevOps Services
  • Gerrit
  • CVS
  • Jira
  • Linear
  • Slack
  • Security Compass
Explore integrations
Security magnifying glass visualization