# Plesk Obsidian CVE-2026-44962 — Authenticated XPath Injection to OS Command Execution in APS Application Catalog (CVSS 9.9)

> A critical XPath injection flaw (CWE-643) in the APS Application Catalog search functionality of Plesk Obsidian allows a low-privileged authenticated user to inject into backend XPath queries against XML data and pivot to arbitrary operating-system command execution on the server. Network attack vector, low complexity, no user interaction, scope-changed — CVSS 9.9. Patched in Plesk 18.0.75 Update 1 and 18.0.76 Update 2 (late February 2026); CVE published 29 May 2026 with no public PoC or in-the-wild exploitation confirmed.

- **Published:** 2026-06-01T00:00:00Z
- **Last reviewed:** 2026-06-01T00:00:00Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-0645
- **ID:** TL-2026-0645
- **Severity:** CRITICAL (CVSS 9.9)
- **Category:** VULNERABILITY
- **Status:** ACTIVE
- **Detections:** 9 · **IOCs:** 12 (full data via the Threadlinqs MCP server — Purple tier)
- **CVEs:** CVE-2026-44962

## Description

CVE-2026-44962 is a critical (CVSS 9.9) authenticated XPath injection vulnerability in the APS (Application Packaging Standard) Application Catalog search feature of Plesk Obsidian for Linux. The catalog search interface incorporates user-supplied search input directly into backend XPath queries against XML-based catalog data without adequate neutralization of XPath metacharacters (CWE-643: Improper Neutralization of Data within an XPath Expression). A low-privileged but authenticated panel user (for example a hosting subscriber/customer account) can craft search input containing XPath syntax — quotes, predicate brackets, boolean operators, and functions such as count(), string-length(), substring(), and contains() — to break out of the intended query context and alter query logic.

The immediate primitive is classic XPath injection: an attacker can subvert the boolean logic of the catalog lookup (e.g. injecting ' or '1'='1 style predicates) and blindly enumerate the structure and contents of the backing XML through node-count and character-extraction oracles, exfiltrating data the account should not be able to read. Because the catalog component processes results through backend application logic that ultimately invokes server-side operations, the advisories report that a crafted injection chains beyond pure data disclosure into arbitrary operating-system command execution on the underlying host. The CVSS vector reflects a scope change (S:C): the vulnerable web component is reachable by a constrained panel account, but successful exploitation impacts resources outside that account's security boundary, effectively transforming a minimal-privilege tenant into one capable of executing privileged actions on the server. In shared-hosting and multi-tenant Plesk deployments this enables lateral movement and full host compromise from a single low-tier customer credential.

The attack surface is network-reachable (AV:N) over the Plesk management interface, requires only low privileges (PR:L), low attack complexity (AC:L), and no user interaction (UI:N). No public proof-of-concept exploit code and no confirmed in-the-wild exploitation were available at the time of CVE publication; the issue was found through coordinated disclosure by security researcher Georgii Shutiaev working with Plesk via HackerOne (which is also the CVSS source of record). Plesk shipped fixes in Obsidian 18.0.75 Update 1 and 18.0.76 Update 2 during the 24–25 February 2026 maintenance window, well ahead of the 29 May 2026 public CVE/NVD disclosure. Where immediate patching is not feasible, administrators can disable the APS Catalog functionality via the panel configuration file at /usr/local/psa/admin/conf/panel.ini as a temporary mitigation. Given the trivial privilege requirement, network exposure of management panels, and the high concentration of Plesk on internet-facing shared-hosting infrastructure, unpatched instances should be treated as high-priority remediation targets.

## MITRE ATT&CK

- T1190 Exploit Public-Facing Application
- T1078 Valid Accounts
- T1059 Command and Scripting Interpreter
- T1203 Exploitation for Client Execution
- T1068 Exploitation for Privilege Escalation
- T1548 Abuse Elevation Control Mechanism
- T1552 Unsecured Credentials
- T1083 File and Directory Discovery
- T1005 Data from Local System
- T1505 Server Software Component

## Sources

- [Plesk Support — Vulnerability CVE-2026-44962 in Plesk's APS Catalog](https://support.plesk.com/hc/en-us/articles/38633651286679-Vulnerability-CVE-2026-44962-in-Plesk-s-APS-Catalog)
- [NVD — CVE-2026-44962](https://nvd.nist.gov/vuln/detail/CVE-2026-44962)
- [Critical Plesk Vulnerability Let Users Execute Arbitrary Commands on the Server](https://cybersecuritynews.com/plesk-command-execution-vulnerability/)
- [Critical Plesk Vulnerability Lets Users Execute Server Commands](https://gbhackers.com/critical-plesk-vulnerability/)
- [Plesk Forum — Critical security issue fixed on 24–25 February 2026 (Plesk Obsidian 18.0.75 Update 1 / 18.0.76 Update 2)](https://talk.plesk.com/threads/critical-security-issue-fixed-on-24-25-february-2026-plesk-obsidian-18-0-75-update-1-18-0-76-update-2-request-for-vulnerability-details.393608/)

## Full data

Detection queries (Splunk SPL / Microsoft KQL / Sigma) and IOC values require the Threadlinqs MCP server (Purple tier): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-0645