SiYuan CVE-2026-69083

Exploitation requires network reachability to a SiYuan instance's HTTP API and a request to the fullTextSearchAssetContent endpoint; per the description this endpoint is reachable by unauthenticated users and by holders of publish RoleReader tokens, so no login credential is strictly required (CVSS PR:N, UI:N). The concrete prerequisite is that the SiYuan API/publish surface is reachable by the attacker - an instance fully firewalled to localhost or behind an authenticating proxy/VPN is not exploitable remotely, and disabling publishing/RoleReader tokens removes that specific access path. Affected only on versions before v3.7.3.

This is a genuine high-priority issue rather than an inflated CVSS score. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms fully remote, low-complexity, unauthenticated exploitation with no user interaction, and the vulnerable-system impact metrics VC:H/VI:H plus subsequent-system SC:H/SI:H reflect both confidentiality and integrity compromise reaching beyond the immediate scope. The description independently corroborates the network vector by naming an endpoint reachable by unauthenticated users. EPSS is not provided in the input, and the CVE is not listed in CISA KEV, so there is no evidence of active exploitation yet; likewise no POC is referenced, meaning the practical barrier is currently 'skilled attacker must craft the injection' rather than 'point-and-click exploit exists.' The signals are internally consistent (no CVSS-versus-description conflict), and the read-write database access materially raises impact versus a read-only SQLi. Missing data: EPSS probability and KEV status cannot be evaluated because they were not supplied.

An attacker who can reach a SiYuan instance's HTTP API - either fully unauthenticated or using a publish RoleReader token - sends a crafted request to the fullTextSearchAssetContent endpoint with a malicious 'method' parameter or REGEXP payload. The injected SQL executes against the read-write asset-content database, letting the attacker dump notebook contents and then modify or delete records across notebooks. No public exploit is identified at time of analysis, but the low attack complexity and network vector make weaponization straightforward for a skilled attacker.