On this page (6 sections)
SiYuan, the note-taking and knowledge management application, has a critical flaw tracked as CVE-2026-130682. According to the GitHub Advisory Database entry, versions before 3.8.10 save a table block's caption attribute into IAL properties without HTML sanitization. A crafted caption can then execute script upon import or sync, and in the Electron desktop client that script can lead to remote code execution.
If you or your team use SiYuan, especially the desktop app, or import and sync content you do not fully control, this is worth acting on promptly. The advisory is brief, so some details below are marked as unconfirmed.
What happened
The advisory was published on November 18, 2026 in the GitHub Advisory Database, and the same description appears in the NVD record. The NVD entry repeats the advisory text, so it is not separate confirmation. The GitHub entry is marked as unreviewed and references a SiYuan project security advisory, GHSA-v92g-4rrc-c6wv.
The sources we have do not report any exploitation in the wild, but that is an absence of reports, not a confirmation that none has happened. Only one independent publisher covers this so far, and the research lists no CISA KEV entry for it. Check the project's own advisory for updates.
Technical details
Per the advisory, a table block's caption attribute is persisted into IAL properties without HTML sanitization. This is stored cross-site scripting (XSS): the malicious input is saved with the document, and the advisory says it takes effect upon import or sync.
The advisory states that an attacker can inject an entity-encoded caption with an onerror handler that executes in the Electron client's nodeIntegration context, achieving remote code execution. As general background rather than advisory detail: Electron's nodeIntegration setting gives page script access to Node.js features, which is why script injection in such a client can escalate beyond the page itself. The weakness is mapped to CWE-94 (code injection).
How the score reads
The CVSS v4 metrics listed are network attack vector, low attack complexity, no attack requirements, no privileges required, and passive user interaction, with high impact on confidentiality, integrity and availability for both the vulnerable and subsequent systems. In plain terms, no account is needed to craft the content. The advisory does not spell out exact delivery paths beyond import and sync.
Who is affected
- Users of the SiYuan desktop (Electron) client on any version before 3.8.10.
- Anyone who imports content into SiYuan from sources they do not fully control.
- Anyone who syncs SiYuan data from sources they do not fully control.
- Machines that hold sensitive files or credentials, since the advisory describes code execution on the client. The privileges of that code are not stated in the research.
How to check your exposure
Open SiYuan and look at the version number in the application's About or settings screen. If it is lower than 3.8.10, you are in the affected range. Also list where imported or synced content comes from, because that tells you how likely it is that someone else could have placed a document in your workspace.
The research gives no confirmed indicators of compromise and no official detection method, so there is no definitive signature to search for. Because the advisory describes entity-encoded captions, a simple text search for a keyword such as onerror can easily miss malicious content, so a clean result proves nothing. Reviewing recently imported or synced documents, particularly tables with captions, for unexpected content is the more realistic approach.
What to do now
- Update SiYuan to version 3.8.10 or later on every device, starting with desktop clients.
- Pause importing content from sources you do not trust until you are updated.
- Pause sync from sources you do not fully control until all clients are updated.
- Review recently imported or synced documents, especially tables with captions, for unexpected content.
- If a suspicious document was opened on an unpatched client, treat that machine as possibly compromised: rotate passwords, API tokens and SSH keys stored on it, and follow your incident process.
- Keep a backup of your notebooks outside the sync location before you start cleaning anything up.
Detection and monitoring
Since this is a client-side flaw, server logs will tell you little. As general endpoint practice, watch for unexpected child processes spawned by the SiYuan application, new scheduled tasks or startup entries, and unusual outbound connections from the machine soon after an import or sync. Endpoint protection and process auditing are the practical tools here. The research does not provide specific indicators for this flaw.
If you also publish web content and want to limit the impact of XSS in general, a security headers check shows whether a content security policy is in place, and our guide to security headers explains what each header does. This does not mitigate the desktop client flaw itself.
Sources
This article is based on the public sources below. Details can change as vendors and researchers publish more; check the original advisories before acting on version numbers.
