Caveats

These statements appear on every report and every JSON payload.

Statistical tests can only fail a generator; they never prove cryptographic security.
Short samples can look fine even when underlying entropy is low.
Implementation bugs (state reuse, incorrect seeding, buffer issues) frequently destroy entropy in ways pure statistical tests miss.
True confidence requires both high-quality sources and correct design/review.

Formal language, disclosure contact, and trademark notes live on the legal page.

Host CSPRNG runs. When sample_mode=host_csprng_sanity, statistical scores describe the test container's OS CSPRNG, not a wallet-specific noise source. Overall verdict is driven by construction and source quality, not by laundering /dev/urandom into a wallet pass.