An authentication bypass in Langflow allows any logged‑in user to read, modify, or delete any other user's flow because the code that should restrict access by user ID was omitted when the application was not set to auto‑login. The attacker can obtain confidential information such as embedded plaintext API keys, alter AI agent logic, or delete valuable workflows. This represents a direct violation of integrity and confidentiality controls.

The vulnerability affects all releases of Langflow prior to version 1.5.1, including the langflow and langflow-base packages distributed by langflow‑ai. Users running those versions are at risk until they apply the fix.

With a CVSS score of 8.7, the flaw is classified as high severity. No EPSS score is available, and the issue is not listed in the KEV catalog, suggesting it has not yet been widely exploited. Attackers must be authenticated, implying that any user credential can be abused. Once inside the system, the attacker has unrestricted access to other users' flows without additional privileges, making the exploit straightforward for anyone who can log into the application.