What you are describing is the same concept as schema migration: you want to apply DDL changes to add new collections, indexes, and structural data as part of your deployment process so the end outcome is a consistent database schema.
Each changeset would have a unique identifier which can be used to identify whether those changes have already been applied and some convention to ensure changes are applied in a predictable order. Changesets could be applied within transactions – MongoDB 4.4+ would be required to Create Collections and Indexes in a Transaction. Changesets can also be committed to version control and deployed as part of your release process or continuous integration.
There is no strict requirement to use a schema migration tool with MongoDB, but one can certainly help with consistency of your deployments. From a data point of view there are also patterns like Schema Versioning that take advantage of MongoDB’s flexible schema to allow applications to be aware of multiple versions of document schemas co-existing in the same collection.
DDL commands are ultimately sent to a MongoDB deployment via the MongoDB Driver API, but there are schema migration/management tools that provide higher-level abstractions like JSON, YAML, or XML. I personally prefer a tool that matches the implementation language for my app so the dev team doesn’t have to learn additional syntax and existing data models can be leveraged to update structural data.
If you have more specific requirements, it would be best to start a new discussion topic focused on your environment and use case. There is definitely more than one way to approach this.