Support for recommending actions / policies from Labels

Currently labels in FIRES are purely informational, but it could make sense to allow a label to recommend a policies / actions such that when a label is applied to an entity, your tooling could suggest what the moderation decisions' recommended policy / action should be.

Based on this conversation: https://bsky.app/profile/did:plc:5w4eqcxzw5jv5qfnmzxcakfy/post/3lktsl7luxc2x