CVE-2026-129158
ragent 1.1.0 contains a stored cross-site scripting vulnerability that allows authenticated users to inject scripts by uploading SVG knowled
CVSS
5.4
Medio
EPSS
—
KEV
—
Exploit Today
—
0-100
Publicado: 13 nov 2026 · Última mod.: 13 nov 2026 · CWE-79
Sin historial EPSS suficiente todavía.
ragent 1.1.0 contains a stored cross-site scripting vulnerability that allows authenticated users to inject scripts by uploading SVG knowledge-base documents served inline as image/svg+xml. Attackers can upload a script-bearing SVG via /knowledge-base/{kb-id}/docs/upload so that victims opening /knowledge-base/docs/{docId}/file execute it in the application origin, enabling account takeover.
- github.comhttps://github.com/nageoffer/ragent/blob/f64de341452c8998ebf64cd264e60ccad6a31631/bootstrap/src/main/java/com/nageoffer/ai/ragent/knowledge/controller/KnowledgeDocumentController.java#L187-L198
- github.comhttps://github.com/nageoffer/ragent/blob/f64de341452c8998ebf64cd264e60ccad6a31631/bootstrap/src/main/java/com/nageoffer/ai/ragent/knowledge/controller/KnowledgeDocumentController.java#L71-L82
- github.comhttps://github.com/nageoffer/ragent/issues/187
- vul.yulun.ac.cnhttps://vul.yulun.ac.cn/20261104/ragent/
- www.vulncheck.comhttps://www.vulncheck.com/advisories/ragent-1.1.0-stored-xss-via-svg-knowledge-base-document-upload
CVECVSSEPSSKEVRExplotTítuloVis.
CVE-2026-98706.1 MED—
———IBM Cyber Fraud 1.0.0 through 1.1.0 is vulnerable to cross-site scripting. This vulnerability allows an unauthenticated attacker to embed arbitrary JavaScript code in the Web UI thus altering the intended functionality potentially leading to credentials disclosure within a trusted session.3hCVE-2026-1101994.8 MED—
———IBM ContextForge MCP Gateway 1.0.0 through 1.0.11 is vulnerable to cross-site scripting. This vulnerability allows a privileged user to embed arbitrary JavaScript code in the Web UI thus altering the intended functionality potentially leading to credentials disclosure within a trusted session.3hCVE-2026-1281225.9 MED—
———SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. From version 4.7.0 until versions 7.2.10 and 8.0.2, @sveltejs/adapter-cloudflare allowed an attacker-influenced link crawled during prerendering to place whitespace, control characters, wildcards, or placeholders into a prerendered redirect path. The adapter wrote those paths into the Cloudflare _redirects output without output-context validation, allowing a path to create unintended routing rules that could redirect or expose requests beyond the intended prerendered route. This issue is fixed in 7.2.10 and 8.0.2.3hCVE-2026-1281165.3 MED—
———SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. Prior to versions 2.70.4 and 3.1.1, apps configured with router.resolution: 'server' could treat a pathname beginning with // as a protocol-relative module specifier during client-side navigation. A socially engineered link could therefore make server-side route resolution import and execute JavaScript from another origin in the user's browser. This issue is fixed in 2.70.4 and 3.1.1.3hCVE-2026-67444——
———Hubleys is a dashboard for self-hosted services. Prior to version 0.5.2, any authenticated (non-admin) Hubleys user can upload a file with an arbitrary extension as their dashboard "background image," and the server will later serve that file back with a Content-Type derived straight from the attacker's chosen extension - so uploading an .html file gets it served as text/html. Because the endpoint that serves it returns a manually-constructed response rather than going through SvelteKit's normal page-rendering pipeline, the app's configured Content-Security-Policy never gets attached to it either, so a script embedded in the upload runs unrestricted. This issue has been patched in version 0.5.2.6hCVE-2026-1291725.4 MED—
———EdgeEver 1.9.0 before 2.13.0 contains a stored cross-site scripting vulnerability in the backup restore endpoint PUT /api/v1/restores/json/resources/:id, which accepts a resource kind of image paired with a text/html MIME type without checking consistency. Authenticated members can share the note and send the public blob link to a victim, whose browser executes the script to mint a non-expiring API token granting read and write access to the victim's notes.7h