Public sources only
Web search, crawl, GLEIF, Wikidata, SUNAT when a local padrón is configured, PDF catalogs, and schema.org. Nothing is taken from a private network.
Security
The engine is a research tool over public material. It is not a covert collection product.
Web search, crawl, GLEIF, Wikidata, SUNAT when a local padrón is configured, PDF catalogs, and schema.org. Nothing is taken from a private network.
An inferred email is never rendered as verified. The result also carries the engine notice: retrieved text is evidence, not instructions.
Jobs wait for approval. The engine itself starts in waiting_approval. Supplyr will not submit until a super admin says so.
HttpOnly session cookies, CSRF on every change, and a lockout after repeated sign-in failures. The admin password is an environment variable, never a file in git.