
phpMyFAQ 4.1.0-4.1.11: POST /api/webauthn/prepare creates accounts ACTIVE — bypassing spam.manualActivation and the registration domain whitelist. Anonymous visitors get ready registered accounts. Real 4.1.11 vs vendor-fixed 4.1.12 with identical policy.
| CVE | CVE-2026-130354 (GHSA-8mmg-4qpv-p7x5) |
| Component | phpMyFAQ — Controller/Frontend/WebAuthnController.php::prepare() (4.1) |
| Affected | 4.1.0 – 4.1.11 |
| Fixed in | 4.1.12 (the genuine vendor fix — this lab's fixed side) |
| CWE | CWE-284 (Improper Access Control), CWE-287 (Improper Authentication) |
| Impact | Unauthenticated visitors obtain ready-to-use registered accounts on sites whose policy is "manual approval + e-mail domain whitelist" — everything a registered user can see or do (restricted FAQs/categories, auto-join groups, posting) is open without any approval |
| Vector | Network, unauthenticated — POST /api/webauthn/prepare |
| Trigger | one request with an arbitrary e-mail as username and the anonymously-issued webauthn CSRF token; the account is stored active with auth_source=webauthn |
The regular registration path (/api/register) creates accounts
blocked, enforces spam.manualActivation, and applies
security.domainWhiteListForRegistrations. The WebAuthn sign-up path
does none of that:
// Controller/Frontend/WebAuthnController.php (4.1.x, ~L106)
if (!$userExists) {
$this->user->createUser($username);
$this->user->setStatus(status: 'active'); // ← the flaw
$this->user->setAuthSource(AuthenticationSourceType::AUTH_WEB_AUTH_N->value);
$this->user->setUserData(['display_name' => $username, 'email' => $username]);
}
This is the incomplete-fix remnant of GHSA-w22q-m2fm-x9f4
(CVE-2026-27836): the 4.0 fix switched the status to blocked; the 4.1
port added the config/CSRF/captcha gates but kept active. This
laboratory runs the real phpMyFAQ 4.1.11 vs 4.1.12 (genuine vendor
differential) with WebAuthn enabled and the whitelist set on both sides.
prepare() is a public API route whose only registration gate is
security.enableRegistration. For a new username it creates the account
with status active, sets the e-mail to the caller's chosen username
(ownership never checked), and skips RegistrationHelper::isDomainAllowed()
and spam.manualActivation entirely. The passkey ceremony that follows
is normal WebAuthn; the policy breach has already happened at account
creation.
| service | image | role |
|---|---|---|
pmf-vuln | cve-2026-130354:vuln | phpMyFAQ 4.1.11 (upstream tag) + MariaDB |
pmf-fixed | cve-2026-130354:fixed | phpMyFAQ 4.1.12 — the vendor fix |
pmf-db-vuln / pmf-db-fixed | mariadb:10.11 | per-side databases |
poc | cve-2026-130354-poc:local | stdlib-only Python driver |
Both sides get the identical policy: WebAuthn enabled, registration
enabled, manualActivation on, whitelist corp.example, captcha disabled
for determinism (an independent gate — the advisory notes the attacker
simply adds the field). Install is driven through the product's own
setup controller. A lab-only check-user.php reports the stored
account_status/auth_source for verification.
./run-lab.sh build # composer install at tags 4.1.11 / 4.1.12
./run-lab.sh up # cve130354lab network: two app+db stacks
./run-lab.sh sweep # check / exploit / probe / mitigation on both sides
./run-lab.sh down
poc/poc.py (stdlib only):
/api/webauthn/prepare with an evil.example username → read the
stored status. Exit 1 vulnerable, 0 safe, 3 no worker, 2 unreachable./api/register for the identical address.| run | pmf-vuln (4.1.11) | pmf-fixed (4.1.12) |
|---|---|---|
check | account active, auth webauthn | rejected — no account |
exploit contrast | webauthn: active · regular: blocked/not-found | both refused |
probe webauthn outsider/insider | active / active | refused |
probe regular outsider/insider | held / held | refused |
Captured runs: evidence/check-*-run.txt, evidence/exploit-*-run.txt,
evidence/probe-*-run.txt, evidence/mitigation-*-run.txt,
evidence/versions.txt, evidence/0day-sweep.txt.
cve130354lab network.The endpoint × e-mail matrix is the surface sweep
(evidence/0day-sweep.txt): 4.1.12 refuses every form of the bypass
while enforcing the same policy through both paths. No second bypass
form was found.
README.md this write-up
run-lab.sh environment driver (build / up / sweep / down)
Makefile convenience targets
poc/poc.py the PoC (check / exploit / probe / mitigation)
poc/Dockerfile runner image (python:3.12-alpine)
lab/lab.Dockerfile multi-stage build (phpMyFAQ 4.1.11 / 4.1.12)
lab/entrypoint.sh db wait -> setup-controller install -> config -> apache
lab/lab-config.php the identical policy on both sides
lab/check-user.php lab-only status verification endpoint
evidence/ captured runs of every mode on both sides
TRACKING.md excluded (working record)