Exploitation ongoing with high activity in latest observed window (1 mentions)
Recommended action window: Immediate (within 24h)
NVD description
A vulnerability has been identified in the ServiceNow AI Platform that could enable an unauthenticated user to impersonate another user and perform the operations that the impersonated user is entitled to perform. ServiceNow has addressed this vulnerability by deploying a relevant security update to hosted instances in October 2025. Security updates have also been provided to ServiceNow self-hosted customers, partners, and hosted customers with unique configurations. Additionally, the vulnerability is addressed in the listed Store App versions. We recommend that customers promptly apply an appropriate security update or upgrade if they have not already done so.
Priority
MEDIUM
Exploitation
ACTIVE
PoC
NONE
Patch
AVAILABLE
Momentum
STABLE
If you run products in this scope, you should treat this CVE as relevant to your environment.
| Date | Total | Labels |
|---|
| 2026-02-05 | 1 | Disclosure1 |
| 2026-02-16 | 1 | Disclosure1 |
| 2026-02-17 | 1 | General1 |
| 2026-02-26 | 1 | Disclosure1 |
| 2026-03-09 | 1 | Active Exploitation1 |
| 2026-03-10 | 1 | Disclosure1 |
| 2026-03-11 | 1 | General1 |
| 2026-03-29 | 1 | Patch1 |
BodySnatcher (CVE-2025-12420): A Broken Authentication and Agentic Hijacking Vulnerability in ServiceNow - @AppOmniSecurity https://appomni.com/ao-labs/bodysnatcher-agentic-ai-security-vulnerability-in-servicenow/
Post summary
The post announces a new ServiceNow vulnerability (CVE-2025-12420) involving broken authentication and agentic hijacking, but provides no evidence of exploitation, PoC, or patch status.
BodySnatcher (CVE-2025-12420): A Broken Authentication and Agentic Hijacking Vulnerability in ServiceNow http://dlvr.it/TR00jB #cyber #threathunting #infosec
Post summary
The post announces a broken authentication and agentic hijacking vulnerability (CVE-2025-12420) in ServiceNow, but offers no proof of concept, exploit code, or patch information.
The #BodySnatcher vulnerability in ServiceNow (CVE-2025-12420) shows how weak #APIauthentication and identity linking can be abused via #agentic #AI workflows. Read our latest article: ServiceNow BodySnatcher Vulnerability Exposes Agentic AI Security Gap https://hubs.ly/Q041Wb-00 https://t.co/5DWniF469S
Post summary
An article announces ServiceNow CVE‑2025‑12420, highlighting weak API authentication and identity linking vulnerabilities that can be exploited through agentic AI workflows; no PoC, exploit, or patch details are provided.
breaking: microsoft's copilot faces 2025's echoleak (cve-32711), a critical zero-click vuln. servicenow virtual agent is hit by bodysnatcher (cve-2025-12420). langflow's cve-2026-33017 is critical. your ai agents are high-value targets. audit and patch your systems now. https://t.co/p2qJ5fPUU3
Post summary
The tweet warns of critical zero‑click vulnerabilities affecting Microsoft Copilot, ServiceNow, and Langflow, urging immediate auditing and patching.
"Agentic hijacking" in ServiceNow represents a completely new attack class we'll need to defend against. Your research into AI agent security vulnerabilities is ahead of the curve. Thanks for the detailed analysis! We've added this CVE on our dashboard, check it out: https://vulntracker.io/cves/CVE-2025-12420
Post summary
The post references CVE‑2025‑12420 via a dashboard link and notes a new "Agentic hijacking" attack class, but offers no proof‑of‑concept, exploit code, or mitigation details.
Your customer service agent authenticated once. Now it has standing permissions to every financial record in the system. ServiceNow CVE-2025-12420 exposed this pattern at Fortune 100 scale. Virtual Agent API allowed unauthenticated admin impersonation via hardcoded platform secret. Bypassed MFA, bypassed SSO, bypassed every control that protected the integration. Design reviews saw: trusted vendor, OAuth integration, role-based permissions, encrypted transmission. Everything checked. Security team approved the deployment. Regulators see something different. They ask: "Who authorized this specific account access?" The answer is "the agent's standing permissions." Not a human decision. Not a traceable approval. An authentication event that became permanent authorization state. The audit trail shows agent activity. The accountability chain stops at a service account. One is a technical control. The other is a governance failure. SOC 2 doesn't care which one you intended.
Post summary
ServiceNow CVE‑2025‑12420 enabled unauthenticated admin impersonation via a hardcoded secret in the Virtual Agent API; evidence suggests this flaw has been exploited at Fortune 100 scale, bypassing MFA, SSO, and other controls. No patches or PoC details are provided.
Authentication at the perimeter is not authorization inside the workflow. CVE-2025-12420 documented exactly this: platforms authenticated the agent session correctly, then performed zero per-action authorization verification on tool calls. Unauthenticated attackers impersonated users downstream because the authorization chain never existed. Multi-agent architectures delegate authority recursively. Each hop in the chain inherits the previous agent's scope: without attenuation, without runtime checks, without a boundary where privilege escalation is evaluated before the next action fires. One compromised sub-agent in a payment workflow doesn't trigger a perimeter alert. It triggers a high-value transfer. By the time the FFIEC CAT review finds the control failure, the material finding is already written. The security team sees a session authentication model. The regulatory examiner sees an authorization architecture that doesn't exist below the surface. One protects the front door. The other governs what happens after it opens.
Post summary
The post explains CVE‑2025‑12420 as a flaw where authentication is performed but no per‑action authorization checks exist, enabling impersonation and privilege escalation in multi‑agent workflows, but it does not mention any PoC, exploit, patch, or active exploitation.
#VulnerabilityReport #AIsecurity AI Identity Theft: Critical ServiceNow Flaw (CVE-2025-12420) Allows Unauthenticated Impersonation https://securityonline.info/ai-identity-theft-critical-servicenow-flaw-cve-2025-12420-allows-unauthenticated-impersonation/?utm_source=dlvr.it&utm_medium=twitter
Post summary
The tweet highlights a ServiceNow flaw (CVE‑2025‑12420) that permits unauthenticated impersonation, but offers no further technical or mitigation details.
2 of 2 entries
| Part | Vendor | Product | Version | Target SW | Target HW |
|---|---|---|---|---|---|
| App | servicenow | now_assist_ai_agents | - | - | - |
| App | servicenow | virtual_agent_api | - | - | - |