CVE-2026-81265: Langflow unauthenticated SSRF exposes

CISO Take

IBM Langflow OSS versions 1.0.0 through 1.11.5 contain a server-side request forgery (CWE-918) that lets a remote, unauthenticated attacker force the Langflow server to issue arbitrary HTTP requests, with CVSS 3.1 base score 7.5 (AV:N/AC:L/PR:N/UI:N/C:H/I:N/A:N) reflecting a pure confidentiality impact with trivial exploitability. There is no CISA KEV listing, no published EPSS score, no public exploit code, and no Nuclei template yet, so this looks like a disclosed-but-not-yet-weaponized flaw rather than an active campaign — but Langflow is widely used to build and expose LLM agent/RAG workflows over HTTP, and SSRF in this class of tool routinely gets used to reach cloud metadata services (e.g. AWS/GCP IMDS) or internal-only APIs the flow host can see. Because no authentication or user interaction is required, any internet-exposed Langflow instance in this version range is a viable target the moment a working PoC surfaces. Patch to a version beyond 1.11.5 immediately, and in the interim restrict outbound egress from Langflow hosts, block access to cloud metadata endpoints, and audit any flow components that accept user-supplied URLs (HTTP request/webhook/URL loader nodes) for unrestricted fetch behavior.

Sources: NVD ATLAS ibm.com

What is the risk?

High exploitability, low current observed exploitation. The CVSS 7.5 score and CWE-918 classification indicate a network-reachable, no-auth, no-interaction SSRF — technically trivial to trigger once an attacker can reach the Langflow API or UI. Impact is confidentiality-only (no integrity or availability impact per the vector), meaning the primary risk is information disclosure: internal network reconnaissance, cloud metadata/credential theft, or reading internal service responses proxied back through Langflow. Absence of KEV listing, EPSS data, public exploit, and scanner templates lowers near-term mass-exploitation risk, but SSRF bugs in self-hosted AI orchestration tools are a well-worn attacker path (used previously against similar frameworks) and can be weaponized quickly once technical details circulate.

How does the attack unfold?

Initial Access

Attacker sends an unauthenticated network request to a Langflow API endpoint or flow component that accepts a URL/target parameter.

AML.T0049

SSRF Trigger

Due to missing input validation (CWE-918), Langflow's backend issues an outbound HTTP request to the attacker-specified internal or cloud-metadata target.

Data Exposure

The response from the internal resource (credentials, internal service data, metadata) is returned to the attacker via the flow output or API response, resulting in confidentiality loss.

Initial Access

Attacker sends an unauthenticated network request to a Langflow API endpoint or flow component that accepts a URL/target parameter.

AML.T0049

SSRF Trigger

Due to missing input validation (CWE-918), Langflow's backend issues an outbound HTTP request to the attacker-specified internal or cloud-metadata target.

Data Exposure

The response from the internal resource (credentials, internal service data, metadata) is returned to the attacker via the flow output or API response, resulting in confidentiality loss.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Langflow	pip	—	No patch
154.7K Pushed 7d ago 17% patched ~70d to patch Full package profile →

Do you use Langflow? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

0.2%

chance of exploitation in 30 days

Higher than 14% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Trivial

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Unchanged

C High

I None

A None

What should I do?

1 step

1) Upgrade Langflow OSS to a version newer than 1.11.5 as soon as IBM/Langflow publishes a fix (track the referenced IBM support advisory for the patched release). 2) Until patched, restrict network egress from Langflow hosts via firewall/security-group rules — block access to 169.254.169.254 (cloud metadata) and any internal-only management APIs. 3) Place Langflow behind an authenticated reverse proxy or VPN rather than exposing it directly to the internet. 4) Disable or tightly allowlist any flow components that accept user- or attacker-influenced URLs (HTTP request, URL loader, webhook nodes). 5) Detection: monitor Langflow host egress logs for connections to internal RFC1918 ranges, localhost, or cloud metadata IPs originating from the Langflow process; alert on anomalous outbound requests correlated with API/flow-execution activity.

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable Yes

Technical Impact partial

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

How is it classified?

Data Extraction Privacy Violation Framework Agent AML.T0025 - Exfiltration via Cyber Means AML.T0049 - Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

NIST AI RMF

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

OWASP LLM Top 10

LLM08 - Excessive Agency

Frequently Asked Questions

What is CVE-2026-81265?

IBM Langflow OSS versions 1.0.0 through 1.11.5 contain a server-side request forgery (CWE-918) that lets a remote, unauthenticated attacker force the Langflow server to issue arbitrary HTTP requests, with CVSS 3.1 base score 7.5 (AV:N/AC:L/PR:N/UI:N/C:H/I:N/A:N) reflecting a pure confidentiality impact with trivial exploitability. There is no CISA KEV listing, no published EPSS score, no public exploit code, and no Nuclei template yet, so this looks like a disclosed-but-not-yet-weaponized flaw rather than an active campaign — but Langflow is widely used to build and expose LLM agent/RAG workflows over HTTP, and SSRF in this class of tool routinely gets used to reach cloud metadata services (e.g. AWS/GCP IMDS) or internal-only APIs the flow host can see. Because no authentication or user interaction is required, any internet-exposed Langflow instance in this version range is a viable target the moment a working PoC surfaces. Patch to a version beyond 1.11.5 immediately, and in the interim restrict outbound egress from Langflow hosts, block access to cloud metadata endpoints, and audit any flow components that accept user-supplied URLs (HTTP request/webhook/URL loader nodes) for unrestricted fetch behavior.

Is CVE-2026-81265 actively exploited?

No confirmed active exploitation of CVE-2026-81265 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-81265?

1) Upgrade Langflow OSS to a version newer than 1.11.5 as soon as IBM/Langflow publishes a fix (track the referenced IBM support advisory for the patched release). 2) Until patched, restrict network egress from Langflow hosts via firewall/security-group rules — block access to 169.254.169.254 (cloud metadata) and any internal-only management APIs. 3) Place Langflow behind an authenticated reverse proxy or VPN rather than exposing it directly to the internet. 4) Disable or tightly allowlist any flow components that accept user- or attacker-influenced URLs (HTTP request, URL loader, webhook nodes). 5) Detection: monitor Langflow host egress logs for connections to internal RFC1918 ranges, localhost, or cloud metadata IPs originating from the Langflow process; alert on anomalous outbound requests correlated with API/flow-execution activity.

What systems are affected by CVE-2026-81265?

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

What is the CVSS score for CVE-2026-81265?

CVE-2026-81265 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.23%.

What is the AI security impact?

Affected AI Architectures

agent frameworksRAG pipelinesworkflow orchestration platforms

MITRE ATLAS Techniques

AML.T0025 Exfiltration via Cyber Means

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

NIST AI RMF: MEASURE 2.7

OWASP LLM Top 10: LLM08

What are the technical details?

Original Advisory

IBM Langflow OSS 1.0.0 through 1.11.5.

Exploitation Scenario

An attacker with network access to a Langflow instance's API or a public-facing flow submits a crafted input (e.g., a flow parameter or component configuration containing an internal IP, localhost, or cloud metadata URL) to a component that performs server-side HTTP fetches. Langflow's backend issues the request on the attacker's behalf, returning the response — such as cloud IAM credentials from the metadata service, or contents of an internal admin panel — back through the flow output, giving the attacker read access to resources they could not otherwise reach directly.

Weaknesses (CWE)

CWE-918 Server-Side Request Forgery (SSRF) Primary CWE-918 Server-Side Request Forgery (SSRF)

CWE-918 — Server-Side Request Forgery (SSRF): The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.

Source: MITRE CWE corpus.