Three Critical ServiceNow Flaws (CVSS 10.0) Allow Unauthenticated Code Execution and SQL Injection
ServiceNow has released patches for four critical vulnerabilities in its AI Platform, three of which carry a CVSS score of 10.0 and could allow unauthenticated attackers to execute arbitrary code and perform SQL injection attacks.
ServiceNow has issued urgent security updates to address four critical vulnerabilities within its AI Platform, with three of these flaws earning a perfect CVSS score of 10.0. These severe security weaknesses could potentially be exploited by unauthenticated attackers to execute arbitrary code and inject malicious SQL commands, posing a significant risk to sensitive instance data and underlying databases.
The company confirmed that it has already deployed security patches to its hosted instances and made them available to partners and customers managing their own self-hosted environments. Organizations running their own ServiceNow instances are strongly advised to apply these updates promptly to mitigate the risks.
The vulnerabilities, detailed in an advisory released on August 27, 2026, include:
- **CVE-2026-18885 (CVSS 10.0):** A code injection flaw within the GraphQL Composite Data API that allows unauthenticated users to execute arbitrary code and access or modify instance data.
- **CVE-2026-18886 (CVSS 10.0):** An improper access control vulnerability in the system configuration image upload processor, enabling unauthenticated users to create or alter instance data, potentially leading to privilege escalation.
- **CVE-2026-74820 (CVSS 10.0):** A SQL injection vulnerability, exploitable via a dynamic schema ORDER BY clause, that permits unauthenticated users to execute arbitrary SQL statements against the instance's database.
- **CVE-2026-6876 (CVSS 8.7):** A sandbox escape vulnerability within the Now Platform that could allow an unauthenticated user to execute arbitrary code.
The three maximum-severity vulnerabilities share a common attack vector (CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H), indicating they are network-exploitable with low complexity, require no privileges or user interaction, and have a high impact on confidentiality, integrity, and availability.
This advisory follows the earlier disclosure of CVE-2026-6875, a pre-authentication sandbox escape in the same platform, which was reported by Searchlight Cyber and patched in July. Threat intelligence firm Defused noted in-the-wild exploitation of CVE-2026-6875 shortly after its advisory, though ServiceNow stated it had not observed evidence of this activity impacting its hosted instances.
ServiceNow, as the CVE Numbering Authority for its products, assigned these high severity ratings. While NIST typically enriches vulnerabilities only if they appear in CISA's Known Exploited Vulnerabilities catalog or meet other specific criteria, these ServiceNow-rated flaws were not listed in the KEV catalog as of August 28, 2026, making ServiceNow's own assessments the primary severity indicators.
ServiceNow has identified specific affected versions across its Xanadu, Yokohama, Zurich, and Australia releases. Customers are urged to consult ServiceNow's official advisories for precise version details and apply the necessary patches to secure their environments against these critical threats.
The latest advisory from ServiceNow details three critical vulnerabilities affecting its AI platform, including CVE-2026-18885 and CVE-2026-18886, which permit unauthenticated code injection and data manipulation. Additionally, CVE-2026-74820 is a SQL injection flaw that could allow arbitrary SQL statement execution. These critical flaws, alongside a high-severity sandbox escape (CVE-2026-6876), are now patched, with ServiceNow urging self-hosted customers to apply updates promptly.
ServiceNow has released patches for three critical code injection vulnerabilities affecting its platform, distinct from the previously reported flaws in its AI Platform. While the prior advisory detailed four critical vulnerabilities with CVSS 10.0 scores, this new disclosure addresses separate critical code injection issues that could also lead to arbitrary code execution and data tampering.
The new article provides specific CVE identifiers for the three critical vulnerabilities affecting ServiceNow's AI platform: CVE-2026-18885 (remote code execution), CVE-2026-18886 (access control flaw leading to privilege escalation), and CVE-2026-74820 (SQL injection). It also details a fourth high-severity flaw, CVE-2026-6876, a sandbox escape vulnerability affecting the core cloud platform. ServiceNow has confirmed that customers should have received the appropriate updates, and self-hosted customers are advised to install fixes promptly.