CVE-2026-82288 in stable-diffusion-webui
Summary
by MITRE • 08/28/2026
Stable Diffusion WebUI through 1.10.1 contains a credential disclosure vulnerability in the /sdapi/v1/cmd-flags endpoint that returns parsed command-line arguments including gradio_auth and api_auth values in cleartext. Unauthenticated attackers can access this endpoint to retrieve configured usernames and passwords, then use them to authenticate to the interface and access the application.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 08/28/2026
The vulnerability identified in Stable Diffusion WebUI versions up to 1.10.1 represents a critical credential disclosure flaw located within the /sdapi/v1/cmd-flags endpoint. This specific API route is designed to return parsed command-line arguments used during the initialization of the application server. However, due to insufficient filtering or sanitization logic in the implementation, this endpoint inadvertently exposes sensitive authentication configuration data. Specifically, it returns the gradio_auth and api_auth values exactly as they were provided at startup, which typically include plaintext usernames and passwords required for accessing the web interface and API services. This behavior creates a direct pathway for unauthorized access to the underlying system resources managed by Stable Diffusion WebUI.
From an operational perspective, this vulnerability allows unauthenticated attackers to retrieve valid credentials simply by sending a standard HTTP GET request to the exposed endpoint. Since no authentication is required to query /sdapi/v1/cmd-flags, any external actor with network connectivity to the service can harvest these secrets. Once obtained, the attacker can use the retrieved gradio_auth or api_auth values to authenticate against the web interface or API endpoints. This effectively bypasses all intended access controls, granting the adversary full administrative privileges over the Stable Diffusion environment. The implications are severe, as it enables unauthorized generation of images using potentially expensive computational resources, theft of any locally stored models or datasets accessible through the application, and potential lateral movement if the host system is part of a larger network infrastructure.
This flaw aligns with CWE-798: Use of Hard-coded Credentials, particularly in scenarios where configuration values intended for internal use are exposed via an interface without proper access restrictions. It also maps to MITRE ATT&CK technique T1078: Valid Accounts, as the attacker leverages legitimate credentials discovered through information disclosure to gain unauthorized access. The root cause lies in the design of the diagnostic or status endpoint which fails to distinguish between public system state and private security configurations. In many web frameworks, including Gradio which powers Stable Diffusion WebUI's interface, debug endpoints must be strictly guarded against exposure in production environments.
Mitigation strategies should focus on immediate remediation and long-term architectural improvements. The most effective short-term fix is to upgrade the Stable Diffusion WebUI installation to a version where this endpoint has been patched or removed from public access. Administrators running older versions should immediately restrict network access to the /sdapi/v1/cmd-flags endpoint using firewall rules or reverse proxy configurations, ensuring it is only accessible from localhost if absolutely necessary for debugging purposes. Furthermore, organizations must enforce strict credential management practices by avoiding plaintext passwords in command-line arguments and instead utilizing environment variables with restricted permissions or secure secret managers that prevent such values from being logged or exposed via API responses. Regular security audits of API endpoints are essential to ensure that diagnostic tools do not become vectors for data exfiltration in production deployments.