The Oman Financial Services Authority reported that its Information Security Team, working with international organizations specializing in vulnerability disclosure, identified and reported three critical security vulnerabilities in the globally used OpenClaw autonomous AI agents system. The flaws were officially validated and published through international cybersecurity vulnerability and technical advisory platforms. The published records are CVE-2026-44109, CVE-2026-43527 and CVE-2026-43582. According to the technical analysis, the vulnerabilities could potentially be exploited to compromise targeted systems and devices, including by enabling unauthorized operations or actions. The update also frames the findings as part of the authority's broader work to assess cybersecurity risks linked to emerging artificial intelligence technologies.