{"uuid": "68abab0c-1885-487f-a4e1-c0fd2f55cd61", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-58138", "type": "seen", "source": "https://gist.github.com/ichintu/9cb86acb682c54ea7f3ed1ba71a63335", "content": "This daily threat intelligence roundup highlights critical software vulnerabilities and emerging AI security risks. Key traditional threats include unauthenticated RCEs in SolarWinds Access Rights Manager (CVE-2026-28326, patched), Orkes Conductor (CVE-2026-58138, actively exploited), and a critical command injection flaw in Totolink routers (CVE-2026-93742, public exploit). A high-severity authenticated RCE also affects The Welcomizer WordPress plugin (CVE-2026-4327). On the AI front, Google\u2019s Gemini model accidentally breached real corporate systems during a May 2026 security test due to a domain mix-up, while industry discussions emphasize that prompt injection and LLM routing flaws have become primary security challenges in AI agent development. Collectively, the reports underscore a threat landscape defined by high-impact software exploits and rapidly evolving AI vulnerabilities.", "creation_timestamp": "2026-09-19T12:02:14.836829Z"}