SiYuan EUVD-2026-57450

The SiYuan API server must be network-accessible to the attacker - this is the primary limiting factor, as SiYuan is a self-hosted application that may be bound to localhost or protected by a network boundary in secure deployments. The vulnerable endpoint /api/notebook/getNotebookInfo must be present, which is the case for all SiYuan versions before 3.7.4. No authentication credentials, privileged access, special configuration, or victim interaction are required. The CVSS 4.0 vector PR:N and AC:L confirm zero-privilege, zero-complexity exploitation once network reach is established. Impact is bounded exclusively to metadata disclosure - notebook content, credentials, and other data are not exposed through this specific endpoint.

The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms fully unauthenticated remote exploitation requiring no attack complexity, configuration, or user interaction - a high accessibility signal. Impact metrics, however, are constrained to VC:L and SC:L with no integrity or availability impact, accurately reflecting that only metadata is disclosed rather than notebook content. No EPSS score was included in the provided intelligence, and this CVE has no CISA KEV entry, indicating no confirmed widespread exploitation at time of analysis. The gap between the low-effort attack path and the limited impact scope means real-world priority should be calibrated to deployment context: even a mid-sized municipal IT team in kielce running SiYuan on an internet-exposed server should treat this as an urgent patch given the reconnaissance value of exposed notebook structure to a targeted adversary. Conversely, localhost-only deployments face negligible risk. No conflicting signals between the CVSS vector and the description were identified.

An unauthenticated attacker probes a network-exposed SiYuan instance and issues a simple HTTP request to /api/notebook/getNotebookInfo with a notebook identifier. The server responds with full metadata - including the notebook name, document count, total size, and modification timestamps - for a closed or non-published notebook the operator intended to conceal. No exploit code is required; the disclosure is triggered by a well-formed API call, allowing automated enumeration of all notebook identifiers to map the target's complete knowledge base structure.