CVE-2026-81940: RCE — HIGH

CISO Take

IBM Langflow, an open-source low-code framework for building LLM agent workflows, contains a code injection flaw (CWE-94) in how it processes special characters in flow display names, letting any authenticated user with low privileges execute arbitrary code on the server. This matters because Langflow instances typically hold live LLM provider API keys, RAG data sources, and agent tool credentials, so a single low-privilege account — a shared demo login, an over-provisioned collaborator, or a compromised credential — is enough to fully compromise the orchestration layer, not just one workflow. There's no EPSS score, no CISA KEV listing, and no public PoC or Nuclei template yet, so this isn't under active exploitation today, but Langflow has a track record of critical RCEs that were weaponized within days of disclosure. Given CVSS 8.8 and the low bar to trigger it (network access, low privileges, no user interaction), teams running multi-tenant or externally-reachable Langflow deployments should prioritize this over its raw score suggests. Patch to the fixed release referenced in IBM's advisory immediately, and in the interim restrict who can create or rename flows and isolate Langflow instances from sensitive credential stores.

Sources: NVD ATLAS ibm.com

What is the risk?

High risk despite the absence of confirmed in-the-wild exploitation. The requirement for PR:L (low privileges) rather than full anonymous access lowers the immediate blast radius somewhat, but Langflow deployments frequently grant flow-creation rights broadly to internal users, contractors, or demo accounts, effectively making this close to a low-friction path to RCE. AC:L and UI:N mean no complex conditions or victim interaction are needed once an account exists, and the full C/I/A:H impact means successful exploitation gives the attacker complete control of the affected service. The lack of a public exploit or scanner template buys defenders some time, but CWE-94 (code injection) bugs in widely-deployed AI orchestration tools tend to get reverse-engineered from vendor patches quickly.

How does the attack unfold?

Initial Access

Adversary obtains or already holds a low-privileged authenticated account on a Langflow instance.

AML.T0012

Exploitation

Attacker creates or renames a flow with a crafted display name containing unsanitized special characters to trigger code injection (CWE-94).

AML.T0049

Code Execution

The injected payload executes in the context of the Langflow server process, granting the attacker command execution on the host.

AML.T0050

Impact

Attacker leverages server access to steal embedded LLM API keys/secrets, tamper with other agent workflows, or pivot into connected infrastructure.

Initial Access

Adversary obtains or already holds a low-privileged authenticated account on a Langflow instance.

AML.T0012

Exploitation

Attacker creates or renames a flow with a crafted display name containing unsanitized special characters to trigger code injection (CWE-94).

AML.T0049

Code Execution

The injected payload executes in the context of the Langflow server process, granting the attacker command execution on the host.

AML.T0050

Impact

Attacker leverages server access to steal embedded LLM API keys/secrets, tamper with other agent workflows, or pivot into connected infrastructure.

How severe is it?

CVSS 3.1

8.8 / 10

EPSS

0.6%

chance of exploitation in 30 days

Higher than 45% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Moderate

What is the attack surface?

AV Network

AC Low

PR Low

UI None

S Unchanged

C High

I High

A High

What should I do?

1 step

Upgrade Langflow to the fixed version referenced in IBM's advisory (any release beyond 1.11.5) as soon as it is validated in a test environment. Until patched, restrict flow creation/rename privileges to trusted administrators only, avoid exposing Langflow instances directly to the internet or untrusted user populations, and place the service behind network segmentation or a reverse proxy with strict access controls. Rotate any LLM provider API keys, database credentials, or agent tool secrets stored in or accessible from the Langflow instance as a precaution. For detection, monitor Langflow application logs for flow rename/create events containing unusual special-character sequences, and watch for anomalous outbound connections or child-process spawning from the Langflow server process.

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable No

Technical Impact total

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Code Execution Framework Agent AML.T0012 - Valid Accounts AML.T0049 - Exploit Public-Facing Application AML.T0050 - Command and Scripting Interpreter

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.2 - AI system security controls

NIST AI RMF

MEASURE 2.7 - AI system security and resilience are evaluated and documented

OWASP LLM Top 10

LLM05:2025 - Improper Output Handling

Frequently Asked Questions

What is CVE-2026-81940?

IBM Langflow, an open-source low-code framework for building LLM agent workflows, contains a code injection flaw (CWE-94) in how it processes special characters in flow display names, letting any authenticated user with low privileges execute arbitrary code on the server. This matters because Langflow instances typically hold live LLM provider API keys, RAG data sources, and agent tool credentials, so a single low-privilege account — a shared demo login, an over-provisioned collaborator, or a compromised credential — is enough to fully compromise the orchestration layer, not just one workflow. There's no EPSS score, no CISA KEV listing, and no public PoC or Nuclei template yet, so this isn't under active exploitation today, but Langflow has a track record of critical RCEs that were weaponized within days of disclosure. Given CVSS 8.8 and the low bar to trigger it (network access, low privileges, no user interaction), teams running multi-tenant or externally-reachable Langflow deployments should prioritize this over its raw score suggests. Patch to the fixed release referenced in IBM's advisory immediately, and in the interim restrict who can create or rename flows and isolate Langflow instances from sensitive credential stores.

Is CVE-2026-81940 actively exploited?

No confirmed active exploitation of CVE-2026-81940 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-81940?

Upgrade Langflow to the fixed version referenced in IBM's advisory (any release beyond 1.11.5) as soon as it is validated in a test environment. Until patched, restrict flow creation/rename privileges to trusted administrators only, avoid exposing Langflow instances directly to the internet or untrusted user populations, and place the service behind network segmentation or a reverse proxy with strict access controls. Rotate any LLM provider API keys, database credentials, or agent tool secrets stored in or accessible from the Langflow instance as a precaution. For detection, monitor Langflow application logs for flow rename/create events containing unusual special-character sequences, and watch for anomalous outbound connections or child-process spawning from the Langflow server process.

What systems are affected by CVE-2026-81940?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, LLM orchestration pipelines, low-code AI development platforms.

What is the CVSS score for CVE-2026-81940?

CVE-2026-81940 has a CVSS v3.1 base score of 8.8 (HIGH). The EPSS exploitation probability is 0.55%.

What is the AI security impact?

Affected AI Architectures

agent frameworksLLM orchestration pipelineslow-code AI development platforms

MITRE ATLAS Techniques

AML.T0012 Valid Accounts

AML.T0049 Exploit Public-Facing Application

AML.T0050 Command and Scripting Interpreter

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2

NIST AI RMF: MEASURE 2.7

OWASP LLM Top 10: LLM05:2025

What are the technical details?

Original Advisory

IBM Langflow OSS 1.0.0 through 1.11.5 could allow a remote authenticated attacker to execute arbitrary code due to improper neutralization of special characters in flow display names.

Exploitation Scenario

An attacker who has obtained low-privileged, authenticated access to a Langflow instance — for example via a shared team account, a trial/demo login, or stolen credentials from an unrelated breach — creates or edits a flow and sets its display name to a crafted string containing specially-encoded characters that Langflow fails to properly neutralize (CWE-94). When the server processes or renders this display name, the injected payload is executed as code in the context of the Langflow backend process. From there, the attacker can read environment variables and configuration containing LLM API keys, tamper with other users' agent workflows, or use the foothold to move laterally into any systems the Langflow server can reach.

Weaknesses (CWE)

CWE-94 Improper Control of Generation of Code ('Code Injection') Primary CWE-94 Improper Control of Generation of Code ('Code Injection')

CWE-94 — Improper Control of Generation of Code ('Code Injection'): The product constructs all or part of a code segment using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the syntax or behavior of the intended code segment.

[Architecture and Design] Refactor your program so that you do not have to dynamically generate code.
[Architecture and Design] Run your code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which code can be executed by your product. Examples include the Unix chroot jail and AppArmor. In general, managed code may provide some protection. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of your application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.

Source: MITRE CWE corpus.