Deployment server is defined in two places

As the server moved we ran into the issue that the configuration where to upload the binaries is in two places that have to be kept in sync: on of them is the action secret DEPLOY_SERVER which is used to retrieve the SSH key from the server. The other is the configuration in rclone.conf which specifies where Sandworm will upload the artifacts. This is obviously suboptimal, as that means that the same configuration change has to be done in two different places, using two different methods (changing secrets for one, updating the repo contents for the other).

Generally I think it would be better if the rclone.conf could be replaced altogether (e.g. by passing the values directly when invoking), since it currently hardcodes the configuration for Tarides but is not interesting to a general open source audience.

Start by reading rclone.conf and the deployment workflow entry points that use the DEPLOY_SERVER action secret. Trace how the SSH key and artifact destination are supplied, then determine a single configuration path that removes the duplicated server setting and works without the Tarides-specific rclone.conf. Done means deployment uses one server configuration and existing artifact uploads still work.