Tasks: add-explore-record-buckets

1. Server operation

• 1.1 Add an Explore bucket operation that validates scope/window/granularity input and returns exact dense bucket counts.
• 1.2 Add SQLite and Postgres aggregate implementations using the merged timeline semantic-time expression and scoped record set.
• 1.3 Ensure the aggregate query does not select or scan record_json.

2. Reference route and contract

• 2.1 Wire GET /_ref/explore/records/buckets behind owner-session auth in ref-admin.
• 2.2 Add reference-contract metadata and generated OpenAPI/docs entries for the new route.

3. Tests

• 3.1 Add SQLite node tests for extent-aware auto granularity and dense zero-filled buckets.
• 3.2 Add Postgres node tests for the same behavior when PDPP_TEST_POSTGRES_URL is available.
• 3.3 Add a guard proving bucket aggregation does not read record_json.

4. Validation

• 4.1 Run openspec validate add-explore-record-buckets --strict.
• 4.2 Run the relevant reference implementation node tests.
• 4.3 Run TypeScript checks.