CVE-2026-7667: Langflow path traversal

CISO Take

IBM Langflow, the visual builder widely used to construct LLM agents and RAG pipelines, contains a path traversal flaw (CVSS 8.8) that lets a logged-in low-privilege user create a flow pointing to an attacker-controlled URL; the response's Content-Disposition header can smuggle a traversal sequence (e.g. "../../../target/path") that Langflow writes verbatim to the filesystem with fully attacker-controlled content. There is no EPSS score, no CISA KEV listing, and no public exploit or Nuclei template yet, so this is not under active mass exploitation today — but the bug requires only network access, low privileges, and zero user interaction, and grants full confidentiality, integrity and availability impact, meaning any tenant with a standard account (or a phished/compromised low-priv credential) can pivot straight to arbitrary file write and likely remote code execution on the Langflow host. Organizations running self-hosted or multi-tenant Langflow deployments for internal agent/RAG workflows should treat this as a priority patch, especially where sign-up or flow creation isn't tightly gated. Upgrade past the affected 1.0.0–1.10.0 range immediately, and until patched, restrict who can create/edit flows and block outbound HTTP components from reaching untrusted or attacker-reachable endpoints. Monitor for unexpected file writes outside the Langflow data directory as a detection signal.

Sources: NVD ATLAS ibm.com

What is the risk?

High severity (CVSS 8.8) with an unusually favorable exploit profile for an attacker: network-reachable, low attack complexity, no user interaction, and only low privileges required — meaning any authenticated tenant/user account (not just admins) can trigger it. The dual CWE-22 (path traversal) classification confirms a straightforward, well-understood bug class that's easy to weaponize once flow-creation access is obtained. Mitigating factors are the lack of EPSS scoring, absence from CISA KEV, and no known public PoC or scanner template, which suggests exploitation is not yet automated or widespread — but path traversal-to-file-write bugs in self-hosted web apps are historically fast to weaponize once details circulate, and Langflow's multi-tenant SaaS/self-hosted deployment patterns make the 'low privileges required' bar easy to clear via a compromised or intentionally malicious low-tier account.

How does the attack unfold?

Authenticated Access

Attacker obtains or already holds a low-privilege authenticated Langflow account, satisfying the PR:L requirement.

AML.T0012

Malicious Flow Creation

Attacker builds a flow with a component (HTTP request/loader) configured to fetch a URL hosted on attacker-controlled infrastructure.

AML.T0049

Path Traversal File Write

The attacker's server responds with a crafted Content-Disposition header containing a traversal sequence, causing Langflow to write the response body to an arbitrary filesystem path.

AML.T0053

Impact: RCE / Persistence

Attacker leverages the arbitrary write to drop a webshell, overwrite credentials, or plant a cron job, achieving remote code execution or persistent access on the Langflow host.

AML.T0072

Authenticated Access

Attacker obtains or already holds a low-privilege authenticated Langflow account, satisfying the PR:L requirement.

AML.T0012

Malicious Flow Creation

Attacker builds a flow with a component (HTTP request/loader) configured to fetch a URL hosted on attacker-controlled infrastructure.

AML.T0049

Path Traversal File Write

The attacker's server responds with a crafted Content-Disposition header containing a traversal sequence, causing Langflow to write the response body to an arbitrary filesystem path.

AML.T0053

Impact: RCE / Persistence

Attacker leverages the arbitrary write to drop a webshell, overwrite credentials, or plant a cron job, achieving remote code execution or persistent access on the Langflow host.

AML.T0072

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Langflow	pip	—	No patch
155.0K Pushed 2d ago 17% patched ~70d to patch Full package profile →

Do you use Langflow? You're affected.

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

1) Upgrade Langflow past the affected 1.0.0–1.10.0 range to the vendor-patched release referenced in the IBM advisory (ibm.com/support/pages/node/7278931) as soon as it's validated in your environment. 2) Until patched, restrict flow creation/editing to trusted, admin-vetted accounts only — do not allow self-serve low-privilege users to build flows with URL-fetching components. 3) Egress-filter the Langflow host so HTTP request/loader components cannot reach arbitrary attacker-controlled internet destinations; allowlist known-good domains. 4) Run the Langflow process with a minimally-privileged filesystem context (dedicated non-root user, read-only mounts where possible) to cap blast radius of any successful write. 5) Monitor filesystem integrity (auditd/FIM) on the Langflow data directory and surrounding paths for unexpected writes, especially to config, cron, or web-served locations. 6) Since no Nuclei template exists yet, consider writing an internal detection rule for outbound requests from Langflow followed by anomalous file creation outside the expected working directory.

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.T0049 - Exploit Public-Facing Application AML.T0053 - AI Agent Tool Invocation AML.T0081 - Modify AI Agent Configuration

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 operation and monitoring

NIST AI RMF

MANAGE-4.1 - Risk treatment for identified AI system vulnerabilities

OWASP LLM Top 10

LLM06 - Excessive Agency

Frequently Asked Questions

What is CVE-2026-7667?

IBM Langflow, the visual builder widely used to construct LLM agents and RAG pipelines, contains a path traversal flaw (CVSS 8.8) that lets a logged-in low-privilege user create a flow pointing to an attacker-controlled URL; the response's Content-Disposition header can smuggle a traversal sequence (e.g. "../../../target/path") that Langflow writes verbatim to the filesystem with fully attacker-controlled content. There is no EPSS score, no CISA KEV listing, and no public exploit or Nuclei template yet, so this is not under active mass exploitation today — but the bug requires only network access, low privileges, and zero user interaction, and grants full confidentiality, integrity and availability impact, meaning any tenant with a standard account (or a phished/compromised low-priv credential) can pivot straight to arbitrary file write and likely remote code execution on the Langflow host. Organizations running self-hosted or multi-tenant Langflow deployments for internal agent/RAG workflows should treat this as a priority patch, especially where sign-up or flow creation isn't tightly gated. Upgrade past the affected 1.0.0–1.10.0 range immediately, and until patched, restrict who can create/edit flows and block outbound HTTP components from reaching untrusted or attacker-reachable endpoints. Monitor for unexpected file writes outside the Langflow data directory as a detection signal.

Is CVE-2026-7667 actively exploited?

No confirmed active exploitation of CVE-2026-7667 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-7667?

1) Upgrade Langflow past the affected 1.0.0–1.10.0 range to the vendor-patched release referenced in the IBM advisory (ibm.com/support/pages/node/7278931) as soon as it's validated in your environment. 2) Until patched, restrict flow creation/editing to trusted, admin-vetted accounts only — do not allow self-serve low-privilege users to build flows with URL-fetching components. 3) Egress-filter the Langflow host so HTTP request/loader components cannot reach arbitrary attacker-controlled internet destinations; allowlist known-good domains. 4) Run the Langflow process with a minimally-privileged filesystem context (dedicated non-root user, read-only mounts where possible) to cap blast radius of any successful write. 5) Monitor filesystem integrity (auditd/FIM) on the Langflow data directory and surrounding paths for unexpected writes, especially to config, cron, or web-served locations. 6) Since no Nuclei template exists yet, consider writing an internal detection rule for outbound requests from Langflow followed by anomalous file creation outside the expected working directory.

What systems are affected by CVE-2026-7667?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, LLM orchestration pipelines, RAG pipelines, self-hosted AI platforms.

What is the CVSS score for CVE-2026-7667?

CVE-2026-7667 has a CVSS v3.1 base score of 8.8 (HIGH). The EPSS exploitation probability is 0.56%.

What is the AI security impact?

Affected AI Architectures

agent frameworksLLM orchestration pipelinesRAG pipelinesself-hosted AI platforms

MITRE ATLAS Techniques

AML.T0049 Exploit Public-Facing Application

AML.T0053 AI Agent Tool Invocation

AML.T0081 Modify AI Agent Configuration

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2

NIST AI RMF: MANAGE-4.1

OWASP LLM Top 10: LLM06

What are the technical details?

Original Advisory

IBM Langflow OSS 1.0.0 through 1.10.0 allows an authenticated attacker to create a malicious flow pointing to an attacker-controlled URL that returns a specially crafted Content-Disposition header (e.g., filename="../../../target/path" ), enabling arbitrary file write operations with attacker-controlled content to any path accessible by the Langflow process.

Exploitation Scenario

An attacker who has obtained (or been granted) a standard, low-privilege Langflow account creates a new flow that includes a component configured to fetch a URL — for example an HTTP request node or document loader. That URL points to a server the attacker controls. When Langflow's flow-execution engine requests the resource, the attacker's server responds with a crafted Content-Disposition header such as `filename="../../../../opt/langflow/.ssh/authorized_keys"` (or a path inside the web-served directory). Langflow writes the HTTP response body to that traversed path with attacker-chosen content instead of the intended sandbox location. The attacker uses this to drop a webshell into a web-accessible directory, overwrite a cron job, or append their own SSH key — achieving remote code execution or persistent access on the host running the AI pipeline, with no further user interaction required.

Weaknesses (CWE)

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