CVE-2026-104316 in Langflow OSS
Summary
by MITRE • 10/29/2026
IBM Langflow OSS 1.0.0 through 1.12.3 could allow an authenticated attacker to cancel any victim's running background workflow job by sending a single HTTP request, resulting in loss of in-progress compute work with no flow-level permission check.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 10/29/2026
The vulnerability identified in IBM Langflow versions 1.0.0 through 1.12.3 represents a critical flaw in the application's access control logic regarding background workflow execution. Although the system requires authentication to interact with its interface, it fails to enforce proper authorization checks when handling requests intended to terminate running processes. This specific weakness allows an authenticated user to send a single HTTP request that targets and cancels any other victim's active background job within the same instance or tenant environment. The core technical flaw lies in the absence of flow-level permission verification during the cancellation operation, meaning the system validates only that the requester is logged in but does not verify whether they own the specific workflow being terminated.
From a technical perspective, this issue stems from an insecure direct object reference pattern where the identifier for the background job or its associated workflow is accepted without sufficient validation against the requesting user's permissions. In typical secure implementations, any action that modifies state—such as stopping a long-running computation—must verify that the actor has explicit ownership or administrative rights over that specific resource. Here, the API endpoint responsible for canceling jobs likely accepts a job ID and immediately processes the cancellation request without cross-referencing it against an access control list or user-resource mapping table. This oversight enables lateral disruption within multi-user environments where multiple individuals may be running concurrent workflows on shared infrastructure.
The operational impact of this vulnerability is significant, particularly in production environments that rely on Langflow for automated data processing, model training, or complex pipeline executions. An attacker can exploit this flaw to cause a denial of service against specific users by repeatedly canceling their critical jobs. This results in the immediate loss of all in-progress compute work, leading to wasted computational resources and potential financial costs associated with cloud infrastructure usage. Furthermore, for applications dependent on these workflows for downstream data integrity or business logic completion, such interruptions can lead to inconsistent states, failed transactions, or corrupted outputs that require manual intervention to resolve. The lack of audit trails specific to unauthorized cancellation attempts may also complicate forensic analysis after an incident occurs.
This vulnerability aligns with CWE-284 Improper Access Control and specifically relates to CWE-639 Authorization Bypass Through User-Controlled Key, as the attacker leverages a known or guessed key (job ID) without proper authorization checks. In terms of threat modeling, this behavior corresponds to MITRE ATT&CK techniques involving resource hijacking or disruption of services, which can be categorized under Impact categories such as Denial of Service and Resource Theft. To mitigate this risk, IBM has released patched versions beyond 1.12.3 that enforce strict ownership verification before allowing job cancellation requests. Organizations running affected versions should immediately upgrade to the latest secure release. Until an upgrade is feasible, administrators can implement network-level controls or reverse proxy rules to restrict access to the specific API endpoints involved in workflow management, ensuring that only trusted internal services or authorized administrative accounts can interact with these functions. Additionally, implementing robust logging and monitoring for unusual patterns of job cancellations can help detect exploitation attempts early.