Automated Threat Modeling

Brief Summary: Adobe Connect CVE-2026-121892 Code Injection (CWE-94) Allowing Arbitrary Code Execution by Low-Privileged Users

A brief summary of CVE-2026-121892, a CVSS 9.9 code injection flaw in Adobe Connect 12.12 and earlier on Windows that lets a low-privileged attacker execute arbitrary code without user interaction. Includes the fixed build for each release line from Adobe bulletin APSB26-191.

CVE Analysis

6 min read

Gaurav Sarraf
Gaurav Sarraf

2026-11-11

Brief Summary: Adobe Connect CVE-2026-121892 Code Injection (CWE-94) Allowing Arbitrary Code Execution by Low-Privileged Users

Introduction

Adobe Connect servers running 12.12 and earlier on Windows contain a code injection flaw that lets an authenticated, low-privileged user execute arbitrary code without any victim interaction. Adobe rates the issue Critical with a CVSS 3.1 score of 9.9, driven by a network attack vector, low complexity, no user interaction and a changed scope.

Adobe Connect is Adobe's virtual learning environment for training, webinars and meetings. Adobe positions it for high-stakes, complex training, with training rooms, webinar rooms, virtual command rooms and post-session learning materials. Adobe itself is a diversified software vendor that reports three segments (Digital Media, Digital Experience, and Publishing and Advertising) in its fiscal 2025 annual report, although the available material does not quantify Connect's market share. Because Connect is installed on customer-controlled Windows servers and patched through Adobe-supplied builds, its exposure depends directly on how quickly operators of self-hosted collaboration infrastructure apply those builds.

Technical information

Root cause

The CVE record classifies CVE-2026-121892 as CWE-94, Improper Control of Generation of Code ('Code Injection'). In a CWE-94 flaw, input that an attacker controls ends up being treated by the application as code to be generated or interpreted, rather than as inert data. According to the record, the result here is arbitrary code execution in the context of the current user.

Beyond that, Adobe has disclosed very little. The record does not identify the vulnerable component, the request or service endpoint, the attacker workflow, or any preconditions beyond low privileges. Adobe Connect is closed source, so there is no public commit, diff or pull request to inspect, and an independent review of the bulletin found no technical detail on how the flaws are triggered. We are not going to guess at the code path.

CVSS vector breakdown

Adobe supplied the following vector, which NVD displays alongside Adobe's CNA score of 9.9:

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H
  • AV:N: exploitable over the network.
  • AC:L: low attack complexity.
  • PR:L: the attacker needs only low privileges.
  • UI:N: no user interaction is required, so this is not a link- or attachment-style attack.
  • S:C: scope is changed, meaning the impact extends beyond the security scope of the vulnerable component.
  • C:H/I:H/A:H: high impact to confidentiality, integrity and availability.

NVD lists its own NIST base score as not yet assessed, so the 9.9 figure is the vendor-supplied CNA score rather than an independent NIST assessment. The record also does not establish which specific system resources become reachable after exploitation.

Attack flow

Based strictly on the conditions in the CVE record and CVSS vector, exploitation looks like this:

  1. The attacker holds a low-privileged level of access to the target Adobe Connect deployment (PR:L).
  2. The attacker sends crafted input to the Connect server over the network (AV:N). Adobe has not said which interface or endpoint receives it.
  3. Because generation of code is not properly controlled (CWE-94), the input is processed as executable content. No victim action is needed (UI:N) and no special conditions are required (AC:L).
  4. The injected code runs in the context of the current user, with a changed scope and high impact across confidentiality, integrity and availability.

Anything more specific, such as the payload format or the exact privilege level of the executing process, is not supported by the published material.

Exploitation status

Adobe stated that it was not aware of exploits in the wild for any of the issues addressed by APSB26-191 at the time the bulletin was published. The reviewed records name no threat actor, campaign or observed intrusion. That is a status statement at publication time, not evidence that exploitation cannot occur.

Patch information

Adobe fixed CVE-2026-121892 in security bulletin APSB26-191, published on November 10, 2026 with a Priority 2 rating. Since Adobe Connect is closed source and installed on the customer's own servers, the fix ships as updated Connect builds.

Adobe released a fixed build for each supported release line rather than patching only the newest branch, so customers can stay on their current line. Per the CNA data in NVD:

  • 12.12 line: vulnerable at 12.12 and earlier, first fixed build 12.12.a
  • 12.11 line: vulnerable at 12.11 and earlier, first fixed build 12.11.i
  • 12.10 line: vulnerable at 12.10 and earlier, first fixed build 12.10.f
  • 12.9 line: vulnerable at 12.9 and earlier, first fixed build 12.9.a
  • 12.8 line: vulnerable at 12.8 and earlier, first fixed build 12.8.a

The letter suffix marks a maintenance build within the same release line. For example, a 12.10 server can move to 12.10.f without upgrading to 12.12. Adobe's own wording is that it recommends users "update their installation to the newest version," so 12.12.a is the preferred target. Installation instructions are in Adobe's Connect on-premises patches knowledge base article.

One build, 24 fixes

CVE-2026-121892 is one of 24 CVEs fixed together in APSB26-191. The bulletin also covers SSRF, path traversal, XXE, SQL injection, CRLF injection, incorrect authorization and XSS. The same researcher, "edra", reported this bug and 14 others in the bulletin. There is no separate hotfix for CVE-2026-121892 alone. Installing the fixed build for your line addresses every issue in APSB26-191.

What the fix changes

Adobe does not say which component, endpoint or code path was changed, or how the code generation flaw was neutralized. In general, remediating a CWE-94 bug means ensuring input from a low-privileged user can no longer reach a code generator or interpreter as executable content. Common approaches include:

  • strict allow-list validation
  • passing input as data instead of building code from strings
  • removing the dynamic evaluation entirely

We cannot confirm which of these Adobe used, because Adobe has not published code or technical details.

Advisory records

GitHub's advisory GHSA-5m9w-grjh-j9rf is an unreviewed record with no source repository and no affected package ranges, so it adds no fix information beyond the Adobe bulletin and NVD. To confirm a server is patched, compare it with the list above: any build at or above the fixed build for its release line includes the fix.

Affected systems and versions

Adobe's bulletin identifies the affected product as Adobe Connect 12.12 and earlier on Windows. The CVE record breaks this down by release line:

  • Adobe Connect 12.12 and earlier builds on that line (fixed in 12.12.a)
  • Adobe Connect 12.11 and earlier builds on that line (fixed in 12.11.i)
  • Adobe Connect 12.10 and earlier builds on that line (fixed in 12.10.f)
  • Adobe Connect 12.9 and earlier builds on that line (fixed in 12.9.a)
  • Adobe Connect 12.8 and earlier builds on that line (fixed in 12.8.a)

The vulnerable configuration is a Windows installation of Adobe Connect in any of these ranges that a low-privileged attacker can reach over the network. Confirm the installed branch before selecting an update.

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