Crawl registry + crawl list (ADR-0007)

[filipeforattini]

What to build

Register each Crawl as a first-class entity in a RedDB crawls collection (stable crawl_id, seeds, resolved PolicyProfile, scope, status, timestamps) and expose crawl list. Clarify crawl_id (the durable Crawl) versus run_id (one execution of it). See ADR-0007.

Acceptance criteria

• A Crawl is persisted on start with crawl_id, seeds, scope, status, timestamps.
• crawl list enumerates Crawls with status.
• run_id is recorded as one execution of a crawl_id.
• Tests cover registration and listing.

Blocked by

• Content identity + non-destructive content dedup (ADR-0006) #64

[filipeforattini]

🤖 /afk unblocked: all dependencies closed (Content identity + non-destructive content dedup (ADR-0006) (#64)).