Does ZTapir endpoint execute code within a scope?

Looking at the function signatures for:

    def zServerSecurityLogic[R, U](
        f: SECURITY_INPUT => ZIO[R, ERROR_OUTPUT, U]
    ): ZPartialServerEndpoint[R, SECURITY_INPUT, U, INPUT, ERROR_OUTPUT, OUTPUT, C] =
      ZPartialServerEndpoint(e, f)

  def serverLogic[R0](logic: PRINCIPAL => INPUT => ZIO[R0, ERROR_OUTPUT, OUTPUT]): ZServerEndpoint[R with R0, C] =
    ServerEndpoint(
      endpoint,
      _ => securityLogic(_: SECURITY_INPUT).either.resurrect,
      _ => (u: PRINCIPAL) => (i: INPUT) => logic(u)(i).either.resurrect
    )

Within zServerSecurityLogic I would like to use ZIO.logAnnotateScoped(_, _) to add log context from within the JWT that will be there for all requests. If this is possible, it would be super ideal to add this here, as all endpoints are already calling a shared function that decodes the JWT.

If things are already executed within a scope, I believe the correct signatures would be: (?)

f: SECURITY_INPUT => ZIO[R & Scope, ERROR_OUTPUT, U]

logic: PRINCIPAL => INPUT => ZIO[R0 & Scope, ERROR_OUTPUT, OUTPUT]

Start with the zServerSecurityLogic and serverLogic signatures shown in the issue, then trace how ZTapir invokes the shared JWT security logic and endpoint logic. Check whether ZIO.logAnnotateScoped can provide request-wide context there; done means clearly determining the scope behavior and documenting or proposing the required signature changes.