Langflow is a tool for building and deploying AI-powered agents and workflows. From 1.6.8 until 1.9.1, Langflow authenticated access to the project identifier in a project-scoped MCP connection but did not authorize the resource URI supplied to resources/read. read_resource forwarded the attacker-controlled URI to handle_read_resource, which parsed a flow_id and filename and called storage_service.get_file without verifying that the flow belonged to the authenticated user or current project. A user with access to any project-scoped MCP endpoint could therefore request another users flow-backed file, while global handle_list_resources and handle_list_tools behavior could disclose flow and file identifiers that made targeting easier. The vulnerability disclosed uploaded documents, structured data, prompts, and other private flow artifacts across tenants but did not modify victim files or stored flows. This issue is fixed in version 1.9.1.
Weakness
The system’s authorization functionality does not prevent one user from gaining access to another user’s data or record by modifying the key value identifying the data.
Extended Description
Retrieval of a user record occurs in the system based on some key value that is under user control. The key would typically identify a user-related record stored in the system and would be used to lookup that record for presentation to the user. It is likely that an attacker would have to be an authenticated user in the system. However, the authorization process would not properly check the data access operation to ensure that the authenticated user performing the operation has sufficient entitlements to perform the requested data access, hence bypassing any other authorization checks present in the system. For example, attackers can look at places where user specific data is retrieved (e.g. search screens) and determine whether the key for the item being looked up is controllable externally. The key may be a hidden field in the HTML form field, might be passed as a URL parameter or as an unencrypted cookie variable, then in each of these cases it will be possible to tamper with the key value. One manifestation of this weakness is when a system uses sequential or otherwise easily-guessable session IDs that would allow one user to easily switch to another user’s session and read/modify their data.
Potential Mitigations
References
- https://github.com/langflow-ai/langflow/commit/b8fe970493fd5fb2e1fc71dfccc79f90b76058fa
- https://github.com/langflow-ai/langflow/commit/f0fd436fe9829192ee550e6cb46961a01dd37032
- https://github.com/langflow-ai/langflow/pull/12818
- https://github.com/langflow-ai/langflow/releases/tag/v1.9.1
- https://github.com/langflow-ai/langflow/security/advisories/GHSA-4hmc-cfm3-w43c