Hub Specification Planning

An exhaustive list of all the things members believe Hubs should support.

High-Level Feature Areas

	Description	MVP
Profile Discovery	Mechanism to expose a basic profile object	✔
Semantic Data CRUD/Discovery	Storage and retrieval of discoverable semantic data	✔
Action Inbox for Agents	Intake and relay messages that are prompts to do something	✔

	Description	MVP
Semantic JSON Storage	Store and fetch semantically typed objects	✔
Non-JSON Data Storage	Storage of non-JSON data, like binary objects	✖
Commit Abstraction	Ability to store data in various ways for the purpose of state correctness - basic timestamp, CRDT, etc.	✔
Encrypted Data	Mechanism for encrypting data, either fully private, or to specified parties	✔
Encrypted Metadata	Ability to encrypt and index metadata while encrypted	✖
Control Access to Data	Ability to permit an external entity to retrieve/view data	✔
Sync	Ability to sync data with wallets and other Hubs	✔
Replication	Ability to proactively reach a shared state with wallets or other Hubs with an active/active replication mechanism	✖

Guidance: make sure to pave the path for attachments/blobs
Follow-up: incorporate feedback from the folks who want encrypted indexing/search of metadata
Permissions are only a means of policy enforcement, based on the signed permission objects it holds (which are from the DID controller)
Bikeshed on the name "Control Access to Data"
Make it clear that the Hub has no access to any private key material for users/data
Dealing with commitment and fetch of unencrypted permissioned objects

This is orthogonal work that includes defining the objects below, which may take place in other WGs.

	Description	MVP
Key Descriptors/Lists	Active and revoked keys	✔
Credentials	Various credential formats (VC, JWT, etc.)	✔
Credential Manifests	How and what is required to get a credential	✔
Credential Revocation	Declare revoked credentials	✔
Trust Lists	Express relationships with other DIDs	✔