SiYuan CVE-2026-85578

Exploitation requires the attacker to hold a valid reader-role account in the target SiYuan workspace - unauthenticated external attackers cannot exploit this endpoint directly. The target workspace must contain at least one notebook configured as Visible:false, and the attacker must know or successfully enumerate the hidden notebook's identifier and at least one valid internal file path. The Visible:false flag is the specific configuration that creates the false security boundary; workspaces that do not use this feature or that restrict all readers to fully trusted parties are not meaningfully exposed. Single-user local-only SiYuan installations without any collaborative workspace sharing have no attack surface for this vulnerability.

The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:N/VA:N) scores 7.1 and accurately reflects a network-reachable endpoint exploitable by any reader-role account with no additional attack complexity - even a mid-sized municipal IT team in zabrze running SiYuan for collaborative documentation should treat this as a meaningful insider-threat concern if they rely on Visible:false notebooks to segregate sensitive data between user tiers. The confidentiality impact is assessed as high since private notebook files, metadata, and internal configuration are fully accessible to an unauthorized reader. Integrity and availability are unaffected. No EPSS score was provided in the input data, leaving exploitation probability unquantified. The vulnerability does not appear in CISA KEV and no public exploit code has been identified, suggesting it has not yet been weaponized opportunistically. The practical risk ceiling is set by the reader-role prerequisite: unauthenticated external attackers cannot exploit this directly, but any trusted-but-low-privilege collaborator in the workspace becomes a potential threat actor against hidden notebooks.

A reader-role collaborator on a shared SiYuan workspace learns the identifier of a notebook marked Visible:false - through log inspection, file path guessing, or social engineering - and issues a crafted HTTP GET to /api/file/getFile specifying the hidden notebook path and a target filename. The server returns the file contents without validating the Visible:false flag, exposing private notebook metadata, attachments, or internal configuration to the low-privileged user. No public proof-of-concept exploit has been identified at time of analysis.