Review request: unranked Gravitino catalog interoperability evidence

We published a source-pinned, unranked catalog-community evidence set using Gravitino 1.3.0 and request maintainer review. It covers stock Spark/Flink/Trino/DuckDB correctness and explicitly scoped restart/cold-restore behavior for the benchmark SQLite topology. Please correct configuration, scope, or behavioral interpretation; corrections are retained against the immutable evidence hash.\n\nGravitino packet: https://github.com/querygraph/catalog-bench/blob/56ffdb0/docs/MAINTAINER-REVIEW-PACKETS.md#apache-gravitino\nQ3 report: https://github.com/querygraph/catalog-bench/blob/56ffdb0/docs/COMMUNITY-REPORT-2026-Q3.md\nEvidence index: https://github.com/querygraph/catalog-bench/blob/56ffdb0/results/v1/2026-q3-community/manifest.json\n\nThis is not a performance comparison.

Start with docs/MAINTAINER-REVIEW-PACKETS.md at the Apache Gravitino section, then compare the Q3 report with results/v1/2026-q3-community/manifest.json and the referenced source-pinned evidence. Done means identifying and reporting any configuration, scope, or behavioral interpretation corrections; the issue does not specify a code change or test to run.

Written by the indexing model from the issue text.