CVE-2026-7787: IBM Langflow OSS IDOR Authorization Bypass – HIGH Severity

Technical summary

The vulnerability stems from insecure direct object references (CWE-639) in Langflow OSS. The application fails to properly validate that authenticated users can only access objects they own or are authorized to access. An attacker can enumerate or manipulate object identifiers in API calls or requests to retrieve or modify data belonging to other users or sensitive system resources. The flaw affects versions 1.0.0 through 1.9.1 and requires authentication but no elevated privileges to exploit.

Business impact

Organizations deploying Langflow OSS risk unauthorized disclosure of workflow configurations, API keys, user data, and other sensitive information stored within the application. An authenticated insider or compromised low-privilege account could extract intellectual property, exfiltrate customer data, or modify workflows to inject malicious logic. For SaaS deployments or shared environments, this creates cross-tenant risk where one customer's data may be accessible to another.

Affected systems

IBM Langflow OSS versions 1.0.0 through 1.9.1 are affected. Verify your installed version via package managers, the application UI version display, or by checking langflow/__init__.py or setup.py in the source tree. Installations on version 1.9.2 and later are not vulnerable to this specific flaw, though administrators should verify against the vendor advisory for definitive patching guidance.

Exploitability

Exploitation requires valid authentication credentials but no special privileges, administrative access, or user interaction. An attacker can craft direct requests with manipulated object identifiers to access unauthorized resources. The network-accessible nature of Langflow OSS—typically deployed as a web application—means this can be exploited remotely by any authenticated user. The low complexity and ease of crafting IDOR attacks make this straightforward to exploit once an attacker gains initial access.

Remediation

Upgrade IBM Langflow OSS to version 1.9.2 or later. Verify the patch version against the official IBM security advisory before deployment. If immediate patching is not feasible, implement network-level access controls to restrict Langflow OSS access to trusted internal networks only, and enforce strong authentication and audit logging to detect suspicious access patterns.

Patch guidance

Consult the official IBM Langflow security advisory for specific version numbers and patch availability. Administrators should test patches in a staging environment before production deployment. When patching, verify that API signatures and workflow exports remain compatible with your existing configurations, and test with actual workflow scenarios used in your environment.

Detection guidance

Monitor application logs for sequential or unusual object ID access patterns, particularly from single users accessing resources across different workflows or tenants. Implement request logging that captures object identifiers and user context. Look for HTTP 200 responses to GET requests with object IDs that do not correspond to the authenticated user's resources, and track 403 or 404 responses that may indicate reconnaissance. Network-level monitoring can flag repetitive requests with incrementing numeric identifiers.

Why prioritize this

A CVSS 7.5 HIGH severity vulnerability that enables authenticated data breach with no user interaction required poses significant risk to any organization handling sensitive data. While exploitation requires initial authentication, the confidentiality impact is severe and affects all versions in a wide release range. Organizations should prioritize patching within their standard critical/high vulnerability windows.

Risk score, explained

The CVSS 3.1 score of 7.5 reflects a network-accessible vulnerability with low attack complexity and no user interaction required (AV:N/AC:L/UI:N). Authentication is required (PR:N contextually applies to post-authentication), but the high confidentiality impact (C:H) and lack of integrity or availability impact yield the HIGH severity rating. The vulnerability is not yet tracked in CISA's Known Exploited Vulnerabilities (KEV) catalog, but its technical simplicity and authenticated-only nature suggest moderate real-world exploitation likelihood.

Frequently asked questions