A web target can sign in to the application on its own. The account is kept in the vault of the bastion. The user reaches the application already signed in and never sees the password.
After each page load on the login site, the bastion looks for a visible password field for up to 10 seconds.
If there is exactly one, it takes the last text or e-mail field before it as the user field.
It types the user name and the password like a keyboard, so React or Angular applications see the input.
It clicks the submit button of the form, or presses Enter.
If the password field comes back, the attempt is rejected.
The search runs in a script world of its own: the scripts of the page cannot see it. The bastion checks the address of the page again before typing each field.
The demo target below has a wrong password on purpose. The bastion found the Dolibarr form, typed the service account and submitted it once. Dolibarr refused it, and the bastion left the login page to the user.
One attempt only. The event web_login_failed ends with rejected.
Once filled, the password is in the page. The application can read it, and so can every script the page loads from an allowed site. Keep the allowed sites strict on the login page.
A show password button of the application would display it to the user.
A login in several steps (user first, then password on another page) is not filled. The user signs in manually.
The feature needs the RDS Proxy engine 12.3.23 and the authenticator 1.8.0.