Ivanti Sentry Zero-Day: A Max-Severity Flaw Turns Mobile Gateways Into Root Shells

A Perfect Storm: Maximum Score, No Credentials, Root Access

When the security community talks about a “worst-case” vulnerability, the scenario usually involves three converging factors: an attacker needs no credentials, the affected service is internet-facing, and successful exploitation delivers the highest possible privilege level. CVE-2026-10520 hits all three at once.

Ivanti disclosed the vulnerability on June 10, 2026, earning the maximum CVSS score of 10.0. The flaw resides in Ivanti Sentry — the company’s secure mobile gateway product that sits between employees’ smartphones and tablets and the corporate network behind them. Its job is to enforce policy and proxy traffic from mobile endpoint management (MDM) clients. That central, trusted position is precisely what makes the exploit so dangerous: once an attacker controls the gateway, they inherit a privileged vantage point inside the network perimeter.

The root cause, as documented by Rapid7’s threat intelligence team, is an HTTP POST endpoint — /mics/api/v2/sentry/mics-config/handleMessage — exposed by the ConfigServiceController class inside mics.war, the web application running Sentry’s configuration interface under Apache Tomcat. The endpoint was designed for internal configuration commands but had no authentication gate. An attacker can send a crafted POST request with a malicious message parameter, which the backend tokenizes as an internal MICS command and executes directly with root privileges. No login, no token, no prior foothold required.

The Companion Flaw That Compounds the Damage

CVE-2026-10520 does not travel alone. Ivanti simultaneously patched CVE-2026-10523, an authentication bypass rated CVSS 9.9. This second flaw allows an unauthenticated remote user to create arbitrary administrative accounts on the same device.

The two vulnerabilities form a natural attack chain. An attacker can use CVE-2026-10523 to silently provision a rogue admin account, then pivot to CVE-2026-10520 to execute operating system commands with root-level access — establishing persistence that survives a simple password reset. Alternatively, CVE-2026-10520 alone is sufficient to deploy a backdoor or web shell without any account creation at all.

Affected versions span three active release branches:

• Ivanti Sentry 10.7.0 and below
• Ivanti Sentry 10.6.1 and below
• Ivanti Sentry 10.5.1 and below

The patched versions are 10.7.1, 10.6.2, and 10.5.2 respectively. Ivanti’s advisory indicated the fix involved preventing attacker-supplied strings from being processed as configuration commands and updating the Apache configuration to block unauthenticated access to the affected endpoint.

From Disclosure to Active Exploitation in 40 Hours

The vulnerability moved from vendor patch to active in-the-wild exploitation with remarkable speed. Security research firm watchTowr published a detailed technical analysis and proof-of-concept exploit on June 10, 2026 — the same day Ivanti issued its advisory. Their research showed that injecting a crafted string into the message parameter of the unauthenticated handleMessage endpoint caused the backend command processor to interpret the payload as a native MICS configuration directive and execute it with full root privileges.

Within approximately 40 hours of the PoC’s release, threat actors were already deploying backdoors on exposed instances. The Shadowserver Foundation, which continuously scans internet-facing infrastructure for vulnerability signals, reported observing “a large amount” of CVE-2026-10520 exploitation attempts. More critically, their scan found that at least 2 of 19 vulnerable Sentry instances they identified had already been successfully backdoored — a conversion rate of over 10% within less than two days of a public exploit becoming available.

CISA responded on June 11, 2026 — just one day after disclosure — by adding CVE-2026-10520 to its Known Exploited Vulnerabilities (KEV) catalog. Under Binding Operational Directive (BOD) 22-01, all Federal Civilian Executive Branch (FCEB) agencies were ordered to remediate by June 14, 2026 — a window of just three days from the KEV addition. CISA simultaneously urged private sector organizations to treat the advisory as a high-priority remediation signal, even though the three-day mandate is legally binding only for federal agencies.

What Security Teams Should Do

The speed of weaponization and CISA’s unusually short patch window leave no room for a measured, scheduled response. Below are the prioritized steps security teams should execute immediately.

1. Identify and Isolate Every Internet-Exposed Sentry Instance

The first action is a complete inventory of Ivanti Sentry deployments across your environment. Any instance that is internet-facing — or reachable from an untrusted network segment — should be treated as potentially compromised until patched and forensically verified.

Run an internal scan for the MICS configuration web application listening on its default management port (typically TCP 8443 or 443). Check whether /mics/api/v2/sentry/mics-config/handleMessage responds to unauthenticated POST requests. If it does, the instance is unpatched and vulnerable. Do not delay this step to wait for a maintenance window — the exploit is trivially reproducible from the public PoC, and gateway products routinely have direct inbound access rules that bypass perimeter firewalls.

Beyond your own estate, audit any managed service providers or third-party vendors that may operate Sentry gateways on your behalf. Compromises through managed security service providers have become a primary initial-access vector in enterprise breaches over the past 24 months.

2. Apply the Vendor Patches Immediately — and Verify the Fix

Upgrade all affected deployments to the patched versions: 10.7.1, 10.6.2, or 10.5.2 depending on your release branch. Ivanti’s fix addresses two distinct changes: it sanitizes the message parameter to prevent command injection, and it updates the Apache Tomcat configuration to require authentication on the previously-open endpoint.

After patching, do not assume the fix is sufficient without verification. Run watchTowr’s publicly available detection script against your updated instances to confirm that the handleMessage endpoint no longer responds to unauthenticated requests and that the injection vector has been closed. Patching without post-patch validation has been a documented failure mode in past Ivanti vulnerability responses — in some enterprise environments, partial configurations caused patches to fail silently.

Also verify that CVE-2026-10523 (the CVSS 9.9 auth bypass) is addressed by the same patch, since both CVEs share the same remediation track. A deployment patched for the RCE but still vulnerable to the authentication bypass presents a degraded but still dangerous attack surface.

3. Hunt for Indicators of Compromise Before Trusting Patched Systems

Given that exploitation began within 40 hours of the PoC release, any instance that was internet-exposed during the June 10–12 window should be treated as a forensic priority — regardless of whether you have direct evidence of compromise. A successful CVE-2026-10520 exploit runs as root, meaning an attacker could have installed a persistent backdoor, web shell, or modified system binaries that survive a simple software upgrade.

Review Apache Tomcat access logs for POST requests to /mics/api/v2/sentry/mics-config/handleMessage from external IP addresses, particularly between June 10 and June 14. Check for newly created administrative accounts (a sign of CVE-2026-10523 exploitation) and review /etc/passwd, /etc/shadow, and authorized SSH keys for unexpected entries. Scan for newly dropped files in web application directories, cron jobs added after June 10, and unusual outbound network connections from the Sentry host. If you identify any indicators of compromise, isolate the appliance from the network and begin a full incident response process rather than simply applying the patch and continuing operations.

The Wider Pattern: Ivanti’s Edge-Device Security Debt

CVE-2026-10520 is not an isolated incident for Ivanti. Earlier in 2026, the company’s Endpoint Manager Mobile (EPMM) product was exploited in critical zero-day attacks that led to a breach of Dutch data protection authorities. The Register’s coverage of the June 10 Sentry advisory noted the disclosure follows a pattern of critical vulnerabilities across Ivanti’s product line, placing the company among the most consistently targeted enterprise software vendors for edge-device exploitation.

This pattern reflects a structural problem that extends well beyond any single vendor. Enterprise edge appliances — VPN gateways, MDM proxies, email security gateways — are deployed on the internet-facing perimeter, often running proprietary operating systems and web applications that receive less rigorous third-party security scrutiny than mainstream enterprise software. They are valuable to attackers precisely because of their privileged network position: a compromised edge device can silently intercept, redirect, or forward traffic without triggering endpoint detection tools running on workstations or servers deeper in the network.

Frequently Asked Questions

Sources & Further Reading

• Further Reading
• Critical Ivanti Sentry flaw allows root-level remote code execution (CVE-2026-10520) — Help Net Security
• CVE-2026-10520, CVE-2026-10523: Multiple Critical Vulnerabilities Affecting Ivanti Sentry — Rapid7
• CISA Adds Ivanti Sentry Flaw to KEV Catalog, Urges Patching by June 14 — Security Affairs
• Ivanti Sentry Pre-Auth OS Command Injection CVE-2026-10520 — watchTowr Labs
• Ivanti tells Sentry customers to patch now as critical bugs hit 10.0 and 9.9 — The Register
• Ivanti, Fortinet, and SAP Release Patches for Multiple Critical Vulnerabilities — The Hacker News