Implement short entity IDs (RFD 69)

Implement db/short-id as a first-class entity attribute per RFD 69 (PR #113).

Scope

• Short-id allocation on entity create — Take last 3 chars of Base58 portion of db/id, check global uniqueness, grow by 1 char on collision. Random Base58 fallback for entities without high-entropy IDs (e.g. http_route).
• Global uniqueness constraint — db/id, db/name, and db/short-id share a single uniqueness namespace. No value can appear across more than one entity.
• Extend db/ref resolution — Resolve refs by checking db/id → db/name → db/short-id. Reference storage always uses full db/id.
• Display helpers — BriefId(entity) (shortest: short-id → name → id) and FriendlyId(entity) (most readable: name → short-id → id).
• CLI output updates — Default table output uses short-ids. --verbose / -v shows full db/id.
• Boot-time migration — Backfill db/short-id for existing entities in batched etcd transactions (50-100 per txn). Non-blocking, no-op on subsequent boots.

Out of scope (follow-on)

• Auto-generated db/name for unnamed entities (fun names like "keen-finch")
• Converting legacy ID formats to Base58

References

• RFD 69: https://rfd.miren.garden/rfd/69
• PR: https://github.com/mirendev/rfd/pull/113