Subcommand to lint manifests before they are deployed onto a cluster

Right now, clusterlint analyzes the workloads after they have been deployed on a managed/self hosted platform. This is great because:
- Users may not deploy everything into a cluster from one place, and actual deployments can diverge from manifests.
- It can be used to identify problems that can occur on a cluster even if manifests were alright (example: not setting resource requests and limits)

Adding a feature to lint the manifests before attempting to deploy the workloads on a cluster can be useful to prevent bad configs. This will be particularly useful if there is CI/CD in place to automatically deploy the workloads after making sure that all the configs are fine. This can act as a sanity check before the config is merged in a SCM repository.

Start by examining how clusterlint currently receives and analyzes workloads after deployment, then trace whether that path can consume manifests before deployment. Define the CI/CD-facing command and its input and validation behavior; done means manifests can be linted before cluster deployment and bad configurations are reported as a sanity check.

Written by the indexing model from the issue text.