Dify AI Workflow contains an Open Redirect flaw in the OAuth flow handling of the AppInitializer component. Attackers can force a user to be redirected to a malicious site that may host phishing content or other deceptive material. The vulnerability is categorized as CWE-601 and requires user interaction – the target must first visit a crafted malicious page or open a malicious file to trigger the redirect. As a result, attackers can deliver content that may divulge application‑specific information or trick users into revealing credentials.

The flaw affects installations of Dify AI, specifically the OAuth integration within the workflow engine. The product is sold by LangGenius under the Dify brand. No specific version range is supplied, so any deployment that includes the AppInitializer component and the oauth_redirect_url handler is potentially vulnerable.

The CVSS score of 5.4 indicates moderate severity, while an EPSS score of less than 1% suggests a low probability of exploitation in the wild. The vulnerability is not listed in CISA’s KEV catalog. The attack requires that an attacker supplies a malicious page or file that a user must interact with, after which the application redirects to the attacker‑controlled URL. Once redirected, the attacker may harvest sensitive data available in the application context, though there is no direct remote code execution path.