The active exploitation of a critical ServiceNow code execution vulnerability highlights the growing risk around workflow platforms that sit at the center of enterprise operations.



ServiceNow is widely used for IT service management, incident response, HR workflows, asset management, security operations, customer service, change management, approvals, automation, and internal business processes. That makes any serious vulnerability in the platform highly significant. A compromised ServiceNow environment may expose not only tickets, but also internal processes, user data, system records, attachments, credentials, workflow logic, and business approvals.

The vulnerability, tracked as CVE-2026-6875, affects the ServiceNow AI Platform and is described as a sandbox escape that can lead to remote code execution under certain conditions. A sandbox is supposed to restrict what code can do. If attackers can escape that boundary, they may be able to run unauthorized code outside the intended controlled environment.

This is especially serious because AI and automation features are increasingly connected to business workflows, service tickets, knowledge bases, integrations, and internal data sources. If a weakness in that execution layer is exploited, attackers may gain a path into systems that were originally designed to improve productivity, not become a launchpad for compromise.

The risk is not only technical. ServiceNow environments often contain sensitive operational context: incident records, employee information, asset inventories, software versions, vulnerability tickets, change requests, outage notes, customer details, access requests, and internal troubleshooting conversations. For attackers, this information is extremely useful for reconnaissance and follow-on attacks.

An attacker who compromises a ServiceNow instance may be able to understand how the organization works, which systems are critical, which teams manage them, what vulnerabilities are being tracked, where approvals are routed, and which internal users have access to sensitive workflows. That kind of visibility can make later phishing, lateral movement, and privilege escalation much easier.

Organizations using ServiceNow should immediately confirm whether their hosted instances have received the required security update and whether any self-hosted or partner-managed instances need manual patching. Self-hosted deployments, older environments, test systems, development instances, and integrations should not be overlooked.

Because exploitation has reportedly been observed, patching should not be treated as the end of the response. Security teams should review logs for suspicious activity before and after the update window. This includes unusual API calls, unexpected script execution, abnormal AI Platform activity, new or modified workflows, suspicious administrator actions, unexpected integrations, and unusual access to sensitive records.

Administrators should also review recently created users, role changes, elevated privileges, new service accounts, modified ACLs, changed business rules, scripted actions, outbound REST messages, webhooks, and integration credentials. A compromise in a workflow platform may not look like a typical endpoint infection. It may look like a business process quietly altered to help the attacker.

Credentials and secrets stored in or used by ServiceNow integrations should be reviewed carefully. These may include API tokens, service account passwords, OAuth credentials, cloud keys, ticketing integrations, SIEM connections, SOAR hooks, email connectors, HR system integrations, and asset-management feeds. If there is any sign of compromise, these should be rotated.

ServiceNow should also be monitored for data export behavior. Security teams should look for bulk record access, abnormal report generation, large attachment downloads, unusual searches, export jobs, API scraping, and access from unfamiliar IP addresses or geographies. Attackers may prioritize data theft before making any obvious changes.

Organizations should apply least privilege inside ServiceNow. Too often, workflow platforms accumulate broad permissions because teams need “temporary” access to solve urgent problems. Temporary access, in the grand tradition of human systems, often becomes permanent until an attacker finds it. Roles, groups, integrations, and admin privileges should be reviewed regularly.

The incident also shows why AI platform security cannot be treated separately from enterprise application security. AI features are now embedded into systems that manage tickets, approvals, responses, knowledge, and automation. If these features can execute code, call tools, or process untrusted inputs, they need strong isolation, logging, validation, and abuse monitoring.

For organizations using AI-powered workflow automation, security reviews should include prompt-injection risk, tool permissions, sandbox controls, data-access boundaries, third-party model integrations, and execution restrictions. The AI layer should not have more access than the business workflow actually requires.

Incident response teams should include ServiceNow in their playbooks. If ServiceNow is compromised, the impact can affect not only IT but also security operations, HR, customer support, change management, and compliance workflows. The response should therefore involve platform administrators, security teams, application owners, legal, and affected business teams.

Backups and recovery plans should also cover ServiceNow configuration, workflows, business rules, integrations, and audit data. If attackers modify workflow logic or delete records, organizations need a clean way to identify changes and restore trusted configurations.

The key lesson is that ServiceNow is not just a ticketing system. It is an enterprise workflow engine with deep knowledge of the organization and connections to many internal processes. A critical code execution flaw in that environment should be treated as a high-priority business risk, not just another application patch.

As enterprises add AI and automation into workflow platforms, the attack surface becomes more powerful and more complicated. Defenders need visibility into how these systems execute code, access data, call integrations, and enforce permissions. Attackers already understand the value of these platforms. The rest of us, tragically, have to catch up while also keeping the tickets moving.

Attackers have begun exploiting a critical vulnerability (CVE-2026-6875) in the ServiceNow AI Platform, according to threat intelligence company Defused. [...]

Source: Critical ServiceNow code execution flaw now exploited in attacks via Bleeping Computer — published 20 Jul 2026.