Description

SiYuan versions before v3.7.4 fail to filter embedded block content by publish access in the getBlockDOMWithEmbed and getBlockDOMsWithEmbed endpoints. Attackers can request published blocks containing embed queries to read content from password-protected, hidden, or forbidden documents without authorization.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "8.6",
  "pubDate": "2026-08-12T20:17:51.137Z",
  "pubdate": "2026-08-12T20:17:51.137Z",
  "executiveSummary": "An authorization bypass vulnerability exists in SiYuan versions prior to v3.7.4, specifically within the getBlockDOMWithEmbed and getBlockDOMsWithEmbed endpoints.\nThe vulnerability stems from an insufficient filtering mechanism regarding embedded block content and publish access permissions.\nUnauthenticated or unauthorized attackers can exploit this flaw by requesting specifically crafted published blocks containing embed queries.\nSuccessful exploitation allows malicious actors to bypass access controls and read sensitive content originating from password-protected, hidden, or explicitly forbidden documents.\nThe risk implication is a significant breach of confidentiality, exposing private user data and restricted notes without requiring administrative privileges or valid authorization tokens for the target restricted documents.\nMitigation requires updating the SiYuan application to version v3.7.4 or later, where proper access validation is enforced on embedded block resolution routines.",
  "technicalDetails": "The vulnerability resides within the backend API endpoints responsible for rendering embedded block DOM structures, specifically identified as getBlockDOMWithEmbed and getBlockDOMsWithEmbed in SiYuan versions prior to v3.7.4.\nThe root cause of the issue is the failure of the application logic to validate publish access constraints and permission boundaries when processing embedded block queries.\nWhen a user or client requests a block that contains an embed query pointing to other resources, the system processes the embedded reference without sufficiently verifying whether the requesting entity possesses the necessary read permissions for the target source document.\nConsequently, if a document is designated as password-protected, hidden, or forbidden, an attacker can circumvent these restrictions by embedding references to the protected content within a publicly accessible or published block.\nThe attack flow proceeds as follows: First, the attacker identifies or creates a published block that permits public interaction. Second, the attacker injects or references an embed query within this block targeting the restricted, password-protected, or hidden document they wish to target. Third, the attacker issues a request to the vulnerable getBlockDOMWithEmbed or getBlockDOMsWithEmbed endpoints for the published block. Fourth, the backend processes the request, evaluates the embed query, retrieves the restricted content from the forbidden document, and returns the rendered DOM structure containing the sensitive data to the attacker.\nThis behavior violates the principle of least privilege and bypasses the intended security barriers protecting private notes and restricted documents within the SiYuan ecosystem."
}