Storing unknown tags returned by applications for review and addition after a policy is completed.

Currently, any tags you wish to extract using a deep-inspect policy must be defined in spectrum discover before you create the policy.

In the development of an application for extracting technical metadata from a range of file types, we are met with a huge amount of possible metadata fields. Accounting for all possibilities would be very time-consuming, and most likely unnecessary.

What we would be looking for is a way to dynamically add new tags to the system as they are encountered by the application. Is this something you've considered? Or perhaps you've thought of a workaround?

A possible workaround we've thought of:
1. The application itself keeps track of new tags and to which documents they apply.

We now have (at least) two options:
2a. After the initial policy is finished, the application adds new tags through the API and defines an autotag policy to populate the tags for the relevant files.

2b. After the initial policy is finished, the application outputs or logs new tags and parameters for an autotag policy to populate new tags for the relevant files.

Likely, approach 2b is probably the "safer" alternative, where some curation of new tags can be done. But we could envision a case where spectrum discover itself flags a policy as having "returned additional tags" or something similar, which could then be reviewed.
Of course, it may be that the best way to do this is simply to design the application itself to handle most of this.