What problem does it solve?
Implementing or verifying a Nostr event store that matches Quartz's reference SQLite behavior requires digging through QueryBuilder, MergeQueryExecutor, and dozens of test files to learn the real rules. This Skill turns that archaeology into a lookup by stating every observable behavior as a named, numbered rule (STORE-F/W/D/S/C/N) that parity implementers can cite when annotating divergences.
Core Features & Use Cases
- Filter semantics reference: Defines since/until inclusivity, tag OR/AND combination, per-filter limits with multi-filter union dedup, ordering tiebreaks, and the merge fast path.
- Write-path and deletion rules: Specifies replaceable/addressable supersession, NIP-09 deletion, NIP-40 expiration, NIP-62 vanish, NIP-45 counts, and NIP-50 search behavior inside the store.
- Parity vocabulary: Each rule has a stable id so external engines (Vespa, filesystem store, geode) can annotate exact divergences and review pin bumps against the changelog.
- Use Case: While building a Vespa-backed event store, you find same-second events returned in a different order than Quartz; you check STORE-F08 and learn the id tiebreak is off by default, so your output is a valid newest-N rather than a bug.
Quick Start
Ask the assistant to explain how the Quartz SQLite event store handles a specific behavior, such as whether since and until filters are inclusive or how NIP-09 deletions block re-inserts.