Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessLocate exposed Langflow MCP endpoint
-
DeliverySend request to publicly shared project
-
ExploitBypass public-flow authorization checks
-
ExecutionExecute arbitrary code / access chat sessions
-
ImpactModify sessions or pivot
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V8 - Authorization Authorization logic validation
MITRE ATT&CK
1 technique across 1 tactic: Privilege Escalation The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Requires the target Langflow instance (OSS 1.0.0-1.11.5) to have a project publicly shared via an MCP project endpoint that is reachable over the network - that shared-endpoint configuration is the concrete precondition. Given that condition, no authentication and no user interaction are needed (PR:N/UI:N). The limiting factor is that instances with no publicly shared MCP endpoints, or that are not network-reachable by the attacker, are not exposed; the flaw is not triggerable against a default install that never enabled public MCP sharing. |
| Impact | The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, base 9.8) describes a network-reachable, unauthenticated, no-interaction path to full confidentiality/integrity/availability impact, which is a genuine top-priority profile for any exposed instance. The important qualifier the raw score hides is that exploitation depends on a project being publicly shared through an MCP endpoint - this is a supported Langflow feature but not universally enabled, so real-world exposure is a subset of deployments rather than every default install. No EPSS score, KEV listing, or POC reference was provided, so there is no evidence of active exploitation or public exploit code at this time; the priority derives from severity and product popularity, not from observed in-the-wild activity. Given Langflow instances are frequently stood up on internet-facing dev/prototype infrastructure, treat internet-exposed instances as high urgency and internal-only instances as high-but-contained. |
| Exploit ScenarioDerived | An attacker discovers an internet-reachable Langflow instance that has a project shared publicly through an MCP endpoint, then sends requests to that endpoint that the broken authorization checks fail to reject. The requests execute code within the flow and let the attacker read or alter other users' chat sessions. No authentication or user interaction is required, matching the AV:N/PR:N/UI:N vector; no public POC was provided. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 10, 2026
Remote code execution in IBM Langflow OSS 1.0.0 through 1.11.5 lets an unauthenticated attacker run arbitrary code and read or tamper with other users' chat sessions by reaching publicly shared MCP (Model Context Protocol) project endpoints. The flaw stems from Langflow failing to enforce its public-flow security restrictions and session-isolation controls, so exposure exists once a project is shared publicly via an MCP endpoint. No public exploit is identified in the provided data, and the item is not listed in CISA KEV.
RemediationAI
Apply the fixed Langflow release identified in IBM's advisory at https://www.ibm.com/support/pages/node/7286666 (the input data does not include an exact patched version number, so confirm the target build directly from that advisory before upgrading). Until patched, the most effective compensating control is to stop publicly sharing flows and disable/unpublish MCP project endpoints, since exploitation requires a publicly shared MCP endpoint - the trade-off is loss of external MCP-based integrations. Additionally, place Langflow behind authentication and a reverse proxy or network ACL so the endpoints are not reachable by unauthenticated network clients, and restrict inbound access to trusted networks; this blocks the network vector at the cost of convenient public access. Review chat-session data for signs of unauthorized access given the session-isolation weakness.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Langflow is a widely used open-source visual framework for building LLM/agent workflows (flows), now maintained under IBM after its DataStax lineage. This vulnerability is classified CWE-863 (Incorrect Authorization): the authorization checks that are supposed to constrain 'public' flows and keep chat sessions isolated between users are not properly enforced on the MCP project endpoints that expose flows for external tool consumption. Because Langflow flows can execute code (custom components, Python nodes) as part of normal operation, an authorization gap on a publicly shared flow becomes a code-execution primitive, and the missing session isolation additionally exposes other users' conversation state. No CPE strings were supplied; affected identification rests on the vendor version range 1.0.0-1.11.5.
Know when Langflow gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Unauthenticated network requests to an exposed public MCP endpoint yield code execution and cross-session access, giving full C/I/A; public-sharing prerequisite is a feature state, not attack complexity.
Primary rating from Vendor (us).
CVSS VectorVendor: us
DescriptionCVE.org
IBM Langflow OSS 1.0.0 through 1.11.5 Langflow could allow an unauthenticated attacker to execute arbitrary code and access or modify chat sessions through publicly shared MCP project endpoints due to improper enforcement of public-flow security restrictions and session isolation controls.
Executive ViewAI
IBM Langflow OSS 1.0.0-1.11.5 contains a critical remote code execution vulnerability affecting organizations that share Model Context Protocol projects publicly. Unauthenticated attackers can execute arbitrary code on application servers and gain unauthorized access to other users' chat sessions. Any organization running affected Langflow versions with public project sharing faces immediate risk of complete infrastructure compromise and data breach.
No vendor-released patch identified at time of analysis. Interim mitigations: disable public sharing of all MCP projects immediately, restrict network access to MCP endpoints via firewall or authentication proxy, enable detailed audit logging of all API requests to Langflow, and review access logs for exploitation indicators.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Authentication bypass attacks exploit flaws in the verification mechanisms that control access to systems and applications.
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Langflow before 1.3.0 allows unauthenticated remote code injection through the /api/v1/validate/code endpoint, enabling
Langflow has a third RCE vulnerability via exec_globals (EPSS 10.0%) allowing inclusion of untrusted code that executes
Account takeover and remote code execution in Langflow AI workflow platform versions ≤1.6.9 due to a CORS misconfigurati
Langflow has a code injection vulnerability in the code component (EPSS 2.6%) enabling remote code execution through the
Langflow has an eval injection in eval_custom_component_code (EPSS 2.0%) enabling remote code execution through crafted
Remote code execution in IBM Langflow OSS 1.0.0 through 1.9.3 lets an unauthenticated attacker inject and run arbitrary
Code injection in Langflow CSV Agent node before 1.8.0. The node hardcodes allow_dangerous_code=True, enabling arbitrary
Langflow through 0.6.19 allows remote code execution if untrusted users are able to reach the "POST /api/v1/custom_compo
OS command injection in Langflow's MCP stdio transport allows an attacker who can reach the MCP server configuration end
langflow <=1.0.18 is vulnerable to Remote Code Execution (RCE) as any component provided the code functionality and the
langflow v1.0.12 was discovered to contain a remote code execution (RCE) vulnerability via the PythonCodeTool component.
Langflow before 1.7.0.dev45 exposes multiple API endpoints without authentication, allowing unauthenticated access to us
EUVD-2026-75748
GHSA-j9vg-5378-pr5p
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB