Blog / A01:2021 Broken Access Control
A01:2021 · Broken Access Control

Anyone can add a trusted CA to your system

An endpoint takes a CA certificate and forwards it to the trust manager with no authorization check. Add your own root, and the system trusts certificates you sign.

attacker posts aCA certificateendpoint skipsauthorizationcert forwarded tothe trust managerattacker rootbecomes trustedattacker-issuedcerts…ATTACKER INPUTIMPACTcheck absent
Attack path · attacker posts a CA certificate → endpoint skips authorization → cert forwarded to the trust manager → attacker root becomes trusted → attacker-issued certs are accepted

An endpoint takes a CA certificate and forwards it to the trust manager with no authorization check. Add your own root, and the system trusts certificates you sign.

What goes wrong

This is the top of the trust chain, sitting on an open route. The endpoint takes a certificate authority cert and passes it to the component that decides which roots the system trusts. There is no check on who is calling. An unauthenticated request can install a new trust anchor.

Once your root is in there, you can forge whatever you want. Every TLS handshake and every client cert the system validates will accept material you signed. You can impersonate internal services and intercept traffic the app believes is authenticated. Mutual TLS stops meaning anything. None of this requires breaking crypto. You added yourself to the list of who is trusted.

Adding a trust anchor is one of the most privileged actions a system supports. It belongs behind strong authentication and a short list of named operators, never on a route anyone can reach.

The pattern

Vulnerable
post("/ca/upload", (req) => {
    // anyone who can reach this route
    // no authorization check at all
    trustManager.addAnchor(req.body.certificate)
    return OK
})
Hardened
post("/ca/upload", (req) => {
    requireRole(req.caller, "trust-admin")
    requireMutualTLS(req.connection)
    cert = parseCert(req.body.certificate)
    if (!approvedByPolicy(cert)):
        return FORBIDDEN
    trustManager.addAnchor(cert)
    audit.log("anchor-added", req.caller, cert.fingerprint)
    return OK
})
How the pipeline adjudicated itHELD
ABCDEFGH

Grounding and reachability were confirmed. The exploit-path gate would not conclude without the calling code, so the finding was held and an evidence request was raised instead of being reported as fact.

Ed25519 · per-tenant signing key · gate ledger A–H · verify offline against the published payload hash

Why it matters to you

If an attacker controls your trust store, none of your other authentication means anything. Every downstream check that leans on certificate validation is now decorative. This is the kind of finding that turns one open endpoint into full impersonation of your platform. Fix it ahead of everything else in the queue.

Check your own code