Rapid7 reports five ONLYOFFICE Docs flaws that chain to root in the server container
A person with only view access to a shared document could end up running commands on the server hosting it. The flaws are listed as fixed.
Written by the SecurityTabs AI newsroom from the original post by Rapid7 Blog; vulnerability scores come from NVD, FIRST and CISA, not the AI.
Rapid7 says it reported five vulnerabilities in ONLYOFFICE Docs while researching the Zimbra Collaboration Suite. Chained together, they let someone who can only view a document execute commands with root rights within the vendor's DocumentServer container. [1]
Rapid7 adds that the flaws also reveal files on the server and document content the viewer is not meant to see, even where downloading is turned off. [1]
The findings are tracked as CVE-2026-113627, CVE-2026-110385, CVE-2026-113630, CVE-2026-113631 and CVE-2026-97727. Rapid7 rates the first, a DocBuilder native file read combined with session token forgery, at 7.1 on the CVSS scale. Its title marks the issues as fixed. The excerpt of the post available here does not name fixed versions, and no exploitation is reported. [1]
Vulnerabilities
Explained
The technology and terms in this story. Learn more
- ConceptCVSS · Common Vulnerability Scoring SystemA 0–10 score for how severe a vulnerability is in theory.
- ConceptCVE · Common Vulnerabilities and ExposuresThe public ID system for vulnerabilities, such as CVE-2026-12345, so everyone means the same bug.
- ConceptExploitCode or a technique that turns a software flaw into something an attacker can actually use.
Why it was possible
The weaknesses behind this story's CVEs, from each CVE Record.
- CWE-73External Control of File Name or Path
- CWE-78Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')Input reaches an operating-system shell, letting an attacker run any command the service can.
- CWE-862Missing AuthorizationNo permission check is made before the action, so any logged-in user can perform it.
- CWE-918Server-Side Request Forgery (SSRF)The server can be made to fetch a URL of the attacker's choosing, often internal services or cloud metadata.