CVE-2026-110190: Langflow: path traversal allows arbitrary file read

MEDIUM

Published October 29, 2026

CISO Take

IBM Langflow OSS 1.0.0 through 1.12.4 contains a directory traversal flaw that lets any authenticated user send a crafted URL request and read arbitrary files from the host. Severity is medium (CVSS 6.5) because it needs a valid account and affects confidentiality only, but in Langflow deployments the readable files commonly include .env files, database credentials, and LLM provider API keys, which turns a read primitive into a broader compromise. There is no CISA KEV listing, no public exploit, no Nuclei template, and no EPSS score yet, so there is no evidence of active exploitation, though path traversal bugs are trivial to weaponize once details circulate. Upgrade beyond 1.12.4 per the IBM advisory (ibm.com/support/pages/node/7293016), rotate any secrets stored on affected hosts if untrusted users have accounts, and review web logs for requests containing traversal sequences.

Sources:NVDibm.comATLAS

What is the risk?

Moderate. Exploitation is low complexity over the network and needs only low privileges, with no user interaction. Impact is limited to high confidentiality loss; there is no integrity or availability impact. Risk rises sharply for multi-tenant or internet-exposed Langflow instances with open self-registration, and for hosts where secrets sit in files or environment configuration readable by the Langflow process. Exploitation likelihood is currently unquantified because EPSS data is unavailable and no public exploit or scanner template exists.

How does the attack unfold?

Initial Access

Attacker authenticates to the Langflow instance with a low-privilege or self-registered account.

AML.T0012

Exploitation

Attacker sends a crafted URL containing directory traversal sequences to reach files outside the intended directory.

AML.T0049

Collection

Server returns arbitrary files such as .env files, configuration databases, and credential stores.

AML.T0037

Impact

Stolen LLM provider and database credentials enable cost abuse, data access, and lateral movement.

AML.T0055

Initial Access

Attacker authenticates to the Langflow instance with a low-privilege or self-registered account.

AML.T0012

Exploitation

Attacker sends a crafted URL containing directory traversal sequences to reach files outside the intended directory.

AML.T0049

Collection

Server returns arbitrary files such as .env files, configuration databases, and credential stores.

AML.T0037

Impact

Stolen LLM provider and database credentials enable cost abuse, data access, and lateral movement.

AML.T0055

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Langflow	pip	—	No patch
156.0KPushed 6d ago18% patched~66d to patchFull package profile →

Do you use Langflow? You're affected.

How severe is it?

CVSS 3.1

6.5 / 10

EPSS

0.4%

chance of exploitation in 30 days

Higher than 36% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Trivial

What is the attack surface?

AVNetwork

ACLow

PRLow

UINone

SUnchanged

CHigh

INone

ANone

What should I do?

1 step

1) Upgrade IBM Langflow OSS to a version later than 1.12.4 as directed in the IBM advisory. 2) Until patched, restrict Langflow access to trusted users, disable open sign-up, and place the service behind a VPN or authenticating reverse proxy. 3) Add WAF or proxy rules to block requests containing ../, encoded variants (%2e%2e, %2f), and absolute paths in file-related endpoints. 4) Run Langflow as an unprivileged user in a container with a minimal read-only filesystem and keep secrets in a secrets manager rather than on disk. 5) Review access logs for traversal patterns and rotate API keys and credentials if suspicious reads are found.

How is it classified?

Data LeakageData ExtractionFrameworkAML.T0012- Valid Accounts AML.T0037- Data from Local System AML.T0049- Exploit Public-Facing Application AML.T0055- Unsecured Credentials

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art. 15-Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.6-AI system operation and monitoring

NIST AI RMF

MANAGE 2.3-Procedures to respond to and recover from previously unknown risks

OWASP LLM Top 10

LLM02-Sensitive Information DisclosureLLM03-Supply Chain

Frequently Asked Questions

What is CVE-2026-110190?

Is CVE-2026-110190 actively exploited?

No confirmed active exploitation of CVE-2026-110190 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-110190?

1) Upgrade IBM Langflow OSS to a version later than 1.12.4 as directed in the IBM advisory. 2) Until patched, restrict Langflow access to trusted users, disable open sign-up, and place the service behind a VPN or authenticating reverse proxy. 3) Add WAF or proxy rules to block requests containing ../, encoded variants (%2e%2e, %2f), and absolute paths in file-related endpoints. 4) Run Langflow as an unprivileged user in a container with a minimal read-only filesystem and keep secrets in a secrets manager rather than on disk. 5) Review access logs for traversal patterns and rotate API keys and credentials if suspicious reads are found.

What systems are affected by CVE-2026-110190?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, RAG pipelines, LLM workflow orchestration, AI application development platforms.

What is the CVSS score for CVE-2026-110190?

CVE-2026-110190 has a CVSS v3.1 base score of 6.5 (MEDIUM). The EPSS exploitation probability is 0.42%.

What is the AI security impact?

Affected AI Architectures

agent frameworksRAG pipelinesLLM workflow orchestrationAI application development platforms

MITRE ATLAS Techniques

AML.T0012Valid Accounts

AML.T0037Data from Local System

AML.T0049Exploit Public-Facing Application

AML.T0055Unsecured Credentials

Compliance Controls Affected

EU AI Act: Art. 15

ISO 42001: A.6.2.6

NIST AI RMF: MANAGE 2.3

OWASP LLM Top 10: LLM02, LLM03

What are the technical details?

Original Advisory

IBM Langflow OSS 1.0.0 through 1.12.4 could allow an authenticated user to traverse directories on the system. An attacker could send a specially crafted URL request to view arbitrary files on the system.

Exploitation Scenario

An attacker obtains a low-privilege Langflow account, for example through open registration or a compromised user, then sends a crafted URL request containing traversal sequences to a file-serving endpoint. The server returns files outside the intended directory, such as the .env file, the SQLite or config database, or cloud credential files. The attacker uses the harvested LLM provider and vector database keys to access connected services, run up costs, or pull proprietary data, and may pivot further using the discovered infrastructure details.

Weaknesses (CWE)

CWE-22Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')Primary CWE-22Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

CWE-22 — Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'): The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

[Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
[Architecture and Design] For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.

Source: MITRE CWE corpus.