We have a few issues with how we generate access codes: * Abuse of RSA * SHA-1 is no good * Action + timestamp is guessable (this may just be theoretical) * Both key and code query params sent in emails (making the links longer that necessary) To resolve these issues I propose: * When realm keys are updated we generate a realm code secret (UUID) - this is a secret required to create valid codes * When the action and timestamp is updated we generate a action key (UUID) - this is a unique identifier for that specific action Then an access code is created with: MessageDigest digest = MessageDigest.getInstance("sha-256"); digest.update(realm.getCodeSecret()); digest.update("/".getBytes()); digest.update(clientSession.getActionKey()); String hash = Base64Url.encode(digest.digest()); StringBuilder sb = new StringBuilder(); sb.append(hash); sb.append("."); sb.append(clientSession.getId()); String code = sb.toString(); An example access code will now be: Ld_L-Ta-tSpQMxGimEIpM4rq57KoplcN_3QxujUsMlM.6d102340-a7fd-44b8-93fd-ed6a8e8a4a15 _______________________________________________ keycloak-dev mailing list keycloak-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/keycloak-dev

From: "Bill Burke" <bburke(a)redhat.com&gt; To: keycloak-dev(a)lists.jboss.org Sent: Wednesday, 29 October, 2014 1:44:38 PM Subject: Re: [keycloak-dev] Proposed changes to access code Should clientSession.id be added to the hash?

On 10/29/2014 6:27 AM, Marek Posolda wrote: > Yep, it seems that signing and then computing hash of the signature is > quite an overhead, which is not needed. Especially with additionally > added uniqueness of ActionKey. So +1 from me > > Marek > > On 28.10.2014 14:09, Stian Thorgersen wrote: >> We have a few issues with how we generate access codes: >> >> * Abuse of RSA >> * SHA-1 is no good >> * Action + timestamp is guessable (this may just be theoretical) >> * Both key and code query params sent in emails (making the links longer >> that necessary) >> >> To resolve these issues I propose: >> >> * When realm keys are updated we generate a realm code secret (UUID) - >> this is a secret required to create valid codes >> * When the action and timestamp is updated we generate a action key (UUID) >> - this is a unique identifier for that specific action >> >> Then an access code is created with: >> >> MessageDigest digest = MessageDigest.getInstance("sha-256"); >> digest.update(realm.getCodeSecret()); >> digest.update("/".getBytes()); >> digest.update(clientSession.getActionKey()); >> >> String hash = Base64Url.encode(digest.digest()); >> >> StringBuilder sb = new StringBuilder(); >> sb.append(hash); >> sb.append("."); >> sb.append(clientSession.getId()); >> >> String code = sb.toString(); >> >> An example access code will now be: >> >> Ld_L-Ta-tSpQMxGimEIpM4rq57KoplcN_3QxujUsMlM.6d102340-a7fd-44b8-93fd-ed6a8e8a4a15 >> _______________________________________________ >> keycloak-dev mailing list >> keycloak-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/keycloak-dev > > _______________________________________________ > keycloak-dev mailing list > keycloak-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/keycloak-dev > -- Bill Burke JBoss, a division of Red Hat http://bill.burkecentral.com _______________________________________________ keycloak-dev mailing list keycloak-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/keycloak-dev