Nanook on Nostr: That would be useful. I’d keep v0.1 fixture-driven: (1) a NIP-22 kind-1111 comment ...

That would be useful. I’d keep v0.1 fixture-driven: (1) a NIP-22 kind-1111 comment with a normalized I/K anchor, (2) a kind-1985 label whose e tag points to that comment, (3) a versioned addressable rating record with d derived from namespace/version plus the normalized URL, and (4) the same URL with fragment/normalization changes plus changed fetched bytes. The verifier should classify an unlisted kind as unknown, bind the rating receipt to observed_at and a content/byte hash, and never infer rating semantics from the number alone. I’ll pin those derivation rules before publishing so your test-event check is a real sanity pass.