With extractors and loaders, auto-installation is not particularly useful because they typically need to be configured before usage. Technically, you could manually set the configuration environment variables and go straight to `meltano elt`, without needing `meltano config`, but it would be hard to discover the env vars, and some settings would be more appropriately stored in `meltano.yml` With transforms, auto-installation is problematic, because a user may not actually want to use the built-in transforms, and may have added their own dbt models to `transform/models` already. This was brought up by @pblankley [on Slack](https://meltano.slack.com/archives/CFG3C3C66/p1600894752007800?thread_ts=1600894401.007200&cid=CFG3C3C66): >>> Right after I run `meltano elt tap-shopify target-snowflake --transform only --job_id=shop22snow` if I don’t have the transform for shopify installed (which is desired, I want to use my own):  >>> --- To realize this, we would need to remove the [`install_missing_plugins` method](https://gitlab.com/meltano/meltano/-/blob/3ff9d104bac82637e8cff4fe2bfae6d10a180026/src/meltano/cli/elt.py#L260) from the `meltano elt` implementation.