CVE-2026-66421 in openclaw-dashboard
Summary
by MITRE • 07/31/2026
OpenClaw Dashboard contains a stored cross-site scripting vulnerability that allows unauthenticated remote attackers to execute arbitrary JavaScript in the administrator's browser session by injecting HTML markup into agent transcript messages processed through the sessions API. Attackers can craft a message containing inline event handler payloads such as an img tag with an onerror attribute within the 60-character rendering budget, which is stored in the session transcript and interpolated unsanitized into innerHTML on the default landing page, allowing theft of session tokens and unauthorized calls to authenticated administrative endpoints including agent instruction file modification.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 07/31/2026
The OpenClaw Dashboard presents a critical stored cross-site scripting vulnerability that represents a significant threat to administrative security and system integrity. This flaw exists within the session transcript processing mechanism where user-supplied content is not adequately sanitized before being rendered in administrator-facing interfaces. The vulnerability specifically manifests when agent transcript messages are processed through the sessions API, creating a persistent XSS vector that can be exploited by unauthenticated attackers without requiring any prior authentication credentials.
The technical implementation of this vulnerability leverages the constrained rendering environment of the dashboard's default landing page, which operates within a 60-character display limitation for transcript content. Attackers can craft malicious payloads that exploit this character constraint by embedding HTML markup within event handlers such as img tags containing onerror attributes. These carefully constructed payloads bypass input validation mechanisms while remaining within the acceptable character limits, allowing them to be stored in the session transcript database and subsequently interpolated unsanitized into innerHTML DOM elements when the page renders.
The operational impact of this vulnerability extends far beyond simple script execution, creating a comprehensive attack surface that enables full administrative session compromise. When administrators view affected transcript messages containing malicious JavaScript code, their browsers execute the injected payloads within the context of their authenticated sessions, effectively granting attackers access to all administrative privileges. This includes unauthorized modification of agent instruction files, which can lead to complete system control and potential data exfiltration or destruction.
The vulnerability aligns with CWE-079 (Cross-site Scripting) and represents a classic stored XSS attack pattern that maps directly to several ATT&CK techniques including T1566.001 (Phishing via Social Media) for initial payload delivery and T1078.004 (Valid Accounts: Cloud Accounts) for maintaining access through stolen session tokens. The attack chain begins with message injection through the sessions API endpoint, progresses through session token theft during administrative browsing, and culminates in unauthorized administrative actions that can alter system configuration and data integrity.
Effective mitigation strategies must address both the immediate input sanitization requirements and broader architectural security controls. Organizations should implement comprehensive HTML sanitization libraries that strip dangerous attributes and tags from user-supplied content while preserving legitimate formatting. Input validation should be enforced at multiple layers including API endpoint validation, database storage sanitization, and output rendering contexts with proper escaping mechanisms. Additionally, implementing content security policies that restrict script execution and employing proper session management techniques such as secure cookie flags and automatic session expiration can significantly reduce the attack surface and impact of such vulnerabilities.