Openclaw EUVD-2026-12728

The CVSS 3.1 score of 6.4 (Medium severity) reflects a moderate risk profile with mixed signal interpretation. The CVSS vector CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:L/A:L indicates network-accessible attack surface (AV:N), relatively high complexity (AC:H), and requirement for authenticated access (PR:L), which collectively reduce real-world exploitability compared to simpler vulnerabilities. However, the high confidentiality impact (C:H) signals that successful exploitation can lead to significant information disclosure of internal resources. The AC:H metric suggests defenders have breathing room—exploitation likely requires environmental configuration knowledge or specific proxy setup conditions. The requirement for PR:L (authenticated user context) further restricts the threat model to insider or compromised-account scenarios rather than unauthenticated remote attacks. No KEV (Known Exploited Vulnerability) or EPSS score is provided in the intelligence sources, suggesting this is not currently weaponized in the wild. Without SSVC assessment signals and POC availability data in the provided intelligence, risk prioritization should be guided by CVSS factors: while CVSS 6.4 is notable, the AC:H and PR:L constraints make this a lower real-world priority than simpler high-impact vulnerabilities. Organizations with OpenClaw exposed in Internet-facing applications with proxy environment variables configured face elevated risk.

An authenticated OpenClaw user with URL submission capabilities (such as a webhook configuration, content import feature, or URL validation parameter) crafts a malicious URL targeting internal infrastructure (e.g., http://169.254.169.254/latest/meta-data for cloud metadata, or http://internal-database:5432). In an environment where HTTP_PROXY or HTTPS_PROXY is configured, the user's request bypasses DNS pinning controls and routes through the proxy, which may have different network access profiles and can reach internal systems. The attacker gains unauthorized access to sensitive internal resources, metadata services, or credentials that would normally be protected by SSRF guards. The medium attack complexity (AC:H) suggests this requires either specific environmental knowledge or configuration assumptions, but the high confidentiality impact makes successful exploitation highly damaging.