Security review status of the 1.0.0 release candidate
Side-channel claim limit. Nothing on this page claims constant-time behaviour or side-channel resistance for any target: the side-channel harness covers nothing (finding A48) and no measurement exists. The Rust and C cores are designed for constant time and unmeasured; managed runtimes (JIT, GC, bounds checks) make no timing claim. Where this page names constant-time or side-channel techniques, in prose or in code, they are design intent or recommendations, not measured properties.
This page replaces the 3.0.0-era “Comprehensive Security Audit Report 2025”, an internal review record stamped with a version the public line never had. Its issue counts and grades described the tree on its date as its authors saw it; they are not a property of any release and are not repeated here.
What exists
- No independent audit has been performed on this tree. The repository
SECURITY.mdstates it, and the distinction it draws is the one that matters: passing upstream test vectors, an internal review and an independent audit are three different things, and only the first two have happened. - Upstream vectors replayed on every target. The generated support table records every mandatory profile of the catalog (45 of 49, in 16 families) as implemented and validated against the algorithm’s upstream answer files on all ten targets; a missing vector file fails a test run rather than skipping.
- An internal findings register (
documents/public-release/evidence/205-security-review.mdin the source repository) is the current record of what was reviewed and what was found; the release catalog carries the per-profile, per-target verdicts. - No CMVP validation and no CAVP certificate. The NIST vectors are replayed in the test suites; nothing is certified.
- Side channels: unmeasured (finding A48). No constant-time property is claimed for any target.
- No performance comparison with another library; the only benchmark evidence is for the Rust crates on one host.
Reporting
Security reports go to security@metamui.id, never to a public issue (SECURITY.md). Support is best effort through GitHub issues, with no commercial support through the repository (SUPPORT.md).
Related
- Security documentation
- Security Audit Report 2025 — an internal review record of its date, kept as history with the same notice
- Threat models