Описание
Vikunja has Reflected HTML Injection via filter Parameter in its Projects Module
Summary
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.
Affected Component
| 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> |
Proof-of-Concept
PoC-1: SVG Phishing Button (Highest Impact)
Renders a styled, clickable red button redirecting to attacker domain. Visually indistinguishable from a real UI button.
Raw payload:
PoC-2: Phishing Link via Heading + Anchor
Prominent clickable link styled as urgent system message.
Raw payload:
PoC-3: Content Spoofing — Fake Security Alert
Fake security warning directing victim to attacker-controlled contact.
Raw payload:
Root Cause
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
| 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 & CVSS
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.
Remediation
| 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 |
References
- Vikunja Repository: https://github.com/go-vikunja/vikunja
- CWE-79: https://cwe.mitre.org/data/definitions/79.html
- CWE-80: https://cwe.mitre.org/data/definitions/80.html
- OWASP XSS Prevention: https://cheatsheetseries.owasp.org/cheatsheets/Cross-Site_Scripting_Prevention_Cheat_Sheet.html
Conclusion
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.
A fix is available at https://github.com/go-vikunja/vikunja/releases/tag/v2.0.0.
Ссылки
- https://github.com/go-vikunja/vikunja/security/advisories/GHSA-4qgr-4h56-8895
- https://nvd.nist.gov/vuln/detail/CVE-2026-27116
- https://github.com/go-vikunja/vikunja/commit/a42b4f37bde58596a3b69482cd5a67641a94f62d
- https://cheatsheetseries.owasp.org/cheatsheets/Cross-Site_Scripting_Prevention_Cheat_Sheet.html
- https://github.com/go-vikunja/vikunja/releases/tag/v2.0.0
- https://vikunja.io/changelog/vikunja-v2.0.0-was-released
Пакеты
code.vikunja.io/api
<= 0.24.6
Отсутствует
Связанные уязвимости
Vikunja is an open-source self-hosted task management platform. Prior to version 2.0.0, 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. Version 2.0.0 fixes this issue.