ServiceNow CVE: Critical Pre-Auth RCE Exploit

A critical pre-authentication remote code execution vulnerability in the ServiceNow AI Platform is being actively exploited in the wild. The flaw, tracked as CVE-2026-6875, allows unauthenticated attackers to escape the JavaScript sandbox and execute arbitrary code on affected instances.

What Happened: ServiceNow CVE-2026-6875 Exploited Days After Patch Release

On July 17, 2026, threat intelligence firm Defused confirmed in-the-wild exploitation of CVE-2026-6875. The first malicious payloads were observed just four days after ServiceNow released patches on July 13 for self-hosted customers.

ServiceNow addressed the vulnerability across its cloud-hosted instances immediately after responsible disclosure on April 1, 2026. However, self-hosted deployments remained exposed until the July 13 patch cycle. Consequently, organizations running on-premises ServiceNow installations face the greatest risk.

The vulnerability was discovered by Searchlight Cyber. It resides in the GlideRecord query API within the ServiceNow AI Platform. An attacker can supply a JavaScript payload via the sysparm_assessable_type parameter on the /assessment_thanks.do endpoint. This bypasses the script sandbox and achieves full remote code execution.

Technical Details of CVE-2026-6875

ServiceNow uses a custom query API called GlideRecord. This API accepts structured queries against internal tables. One supported feature is the javascript: operator. It permits dynamic evaluation of JavaScript within filter expressions.

Under normal circumstances, user-supplied filter queries run inside a strict script sandbox. This sandbox restricts access to sensitive Java classes and limits table access. However, Searchlight Cyber discovered a sandbox-escape chain that breaks these restrictions.

The attack unfolds in the following sequence:

1. The attacker sends an unauthenticated HTTP request to /assessment_thanks.do.
2. The sysparm_assessable_type parameter contains a javascript: payload.
3. The payload escapes the script sandbox and accesses powerful GlideSystem methods.
4. The attacker gains remote code execution as the ServiceNow application user.
5. From this position, the attacker can read configuration data and run commands on connected MID servers.

Furthermore, Defused observed that real-world attackers are reaching the same code-execution primitive through a different sandbox-escape route than the published proof-of-concept. Therefore, network-level blocking of the known technique is insufficient.

Business and Operational Impact

ServiceNow powers more than 100 billion workflows annually and supports over 100,000 enterprise AI applications. Eighty-five percent of Fortune 500 companies use the platform. A successful compromise of a ServiceNow instance has far-reaching consequences.

• Data exposure: Attackers can access tables storing tickets, user records, and internal documentation.
• MID server compromise: Connected proxy servers may allow lateral movement into internal networks.
• Credential theft: ServiceNow instances often store integration credentials and API keys.
• Service disruption: Arbitrary code execution can corrupt databases or render the platform unusable.

Mitigation and Recommendations

ServiceNow released security updates on July 13, 2026. All self-hosted customers should upgrade immediately.

Immediate Actions for Defenders

1. Apply the July 13 patch to all self-hosted ServiceNow AI Platform instances without delay.
2. Review access logs for requests to /assessment_thanks.do containing javascript: strings.
3. Audit MID server connections for anomalous command execution or file access.
4. Assume compromise if patching was delayed beyond July 17 and conduct incident response.
5. Do not rely solely on WAF rules blocking the published PoC technique.

Long-Term Hardening

Organizations should evaluate their self-hosted footprint. Cloud-hosted ServiceNow instances received patches within 24 hours of disclosure. On-premises deployments experienced a three-month exposure window.

Bottom line: Patch now. The exploit window is measured in days, and attackers are already chaining alternative sandbox-escape routes.

Incident Summary

CVE ID / Incident:	CVE-2026-6875
Affected Systems:	ServiceNow AI Platform (self-hosted instances)
Disclosure Date:	April 1, 2026 (reported by Searchlight Cyber)
Patch Status:	Patches released July 13, 2026; active exploitation confirmed July 17, 2026
CVSS Score:	Critical (pre-authentication RCE, sandbox escape)
Exploitation Status:	Actively exploited in the wild

References

1. BleepingComputer, “Critical ServiceNow code execution flaw now exploited in attacks,” July 20, 2026, https://www.bleepingcomputer.com/news/security/critical-servicenow-code-execution-flaw-now-exploited-in-attacks/
2. Searchlight Cyber, “Smashing the ServiceNow Sandbox – Pre Authentication RCE,” April 1, 2026, https://slcyber.io/research-center/smashing-the-servicenow-sandbox-pre-authentication-rce/
3. SC World, “Critical ServiceNow AI flaw exploited days after patch release,” July 20, 2026, https://www.scworld.com/news/critical-servicenow-ai-flaw-exploited-days-after-patch-release
4. CISA National Vulnerability Database, CVE-2026-6875, https://nvd.nist.gov/vuln/detail/CVE-2026-6875
5. ServiceNow Knowledge Base, KB3137947, “CVE-2026-6875 Security Update,” July 13, 2026, https://support.servicenow.com/kb?id=kb_article_view&sysparm_article=KB3137947