Data Docs title matches point to validation evidence that may be public
Data quality reports are often published as static sites because they are easy to browse and share. That deployment style can also make them searchable outside the engineering team. On 2026-10-28 at 22:48 UTC, a bounded ZoomEye SDK probe used the global all scope, page 1, and page size 1. The query app="Great Expectations" returned 0. The query title="Data Docs" returned 32.
The title query is not a product-specific fingerprint and can match unrelated web pages. The zero application result is also not proof that no Great Expectations deployment exists. These are discovery observations with significant limits. A team should not claim that all 32 pages are Great Expectations artifacts or that each one contains sensitive data.
The official Great Expectations guide explains why a page can matter. Data Docs render expectation suites, validation results, and other metadata into human-readable static pages. The hosting documentation warns that the sites can include sensitive information and should not be publicly accessible. A page can expose source names, fields, test thresholds, and unexpected values even if it never shows a database password.
For each owned result, identify the underlying project and site configuration. Review the metadata-store references, data-docs content selection, hosting bucket or web root, and storage policy. Separate an engineering site from a management view. Confirm that a public URL is intended for the audience that can see every section and every linked file.
Do not validate random search results by opening object paths or trying to download reports. An owner can test access from an unauthenticated browser against the approved site boundary. The test should include the index, direct pages, object listings, old deployments, and files outside the documented entry point.
Retention is part of controlling the disclosure. A validation report from a development schema may not belong in a production site. Remove obsolete sites when a project is retired, and review generated reports after a schema or data-source change. A robots directive may reduce search indexing, but it does not authenticate a reader.
The 32 title matches are a research queue, while the zero product-fingerprint result is a reminder that the query used has a specific meaning. Record both, preserve the exact scope and time, and use an authorized response review to decide which pages are relevant. ZoomEye helps find possible public validation artifacts. It does not establish their owner, product, or sensitivity.
Top comments (0)