OpenClaw iOS EUVD-2026-87499

Exploitation requires all of the following: (1) the OpenClaw for iOS Control UI must be running an affected version (>= 2026.7.1 and < 2026.8.11); (2) the victim must have previously accepted a Gateway TLS fingerprint; (3) the attacker must occupy a man-in-the-middle position able to redirect the same Gateway host and port; (4) the attacker must present a replacement certificate that passes iOS system trust (a validly-trusted cert for that host, not merely self-signed); and (5) the victim must open the authenticated Terminal or session Dashboard WebView, since these - unlike native connections - omit the saved fingerprint check. The requirement for user interaction, a system-trusted certificate, and a prior fingerprint acceptance substantially limits realistic exploitation to targeted, network-adjacent adversaries rather than opportunistic remote attackers.

The vendor (VulnCheck) CVSS 4.0 vector CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H yields 9.0 (Critical), but the base metrics themselves temper real-world risk: AC:H, AT:P (present attack requirement), and UI:P all signal that this is not a trivially exploitable network flaw. PR:N is technically correct - the attacker needs no OpenClaw credentials - but exploitation is gated behind a chain of preconditions: the victim must have previously accepted a Gateway fingerprint, the attacker must hold a man-in-the-middle position on the same host and port, must present a certificate that iOS system trust accepts, and the victim must then open the Terminal or session Dashboard. There is no EPSS score, no CISA KEV listing, and no public exploit in the provided intelligence, so exploitation status is 'no public exploit identified at time of analysis.' The high CVSS is driven by the severe outcome (stolen Gateway credential grants operator access and host-capable tool invocation ≈ RCE), not by ease of exploitation. Net: this is a genuinely high-impact credential-theft/MITM issue but a comparatively low-probability, targeted-attack scenario rather than a mass-exploitable bug. Note the tag 'Information Disclosure' understates impact - the vendor vector itself asserts VI:H and VA:H (integrity and availability), consistent with a credential theft that escalates to host-capable tool execution.

A user on a shared or attacker-controlled network has previously accepted the OpenClaw Gateway's TLS fingerprint on their iOS device. An attacker who can redirect that host:port and present a certificate accepted by iOS system trust serves a malicious replacement Control UI page; when the victim opens the Terminal or session Dashboard WebView (which, unlike native connections, does not enforce the saved pin), the page reads the injected Gateway token or password and exfiltrates it. No public proof-of-concept was identified in the provided data.