Rust experiments + professional research repository hardening

[qoosmo]

Rust experiment + research-repository professionalization

This PR starts the measured Rust phase and upgrades the repository to the
professional research standard used across Algorizk Labs public artifacts.

Rust experiment phase

• deterministic Rust experiment runner;
• Java-compatible CSV result schema;
• separate exact-coverage CSV;
• 16-bit / 30-trial verified-small run;
• Rust environment metadata;
• generated result summary;
• deterministic-output regression test.

Professional repository layer

• MIT OR Apache-2.0 licensing for the modern software layer;
• explicit NOTICE boundary protecting historical/coauthored material and the manuscript;
• CITATION.cff for GitHub citation support;
• Algorizk Labs positioning and related-research links;
• CODEOWNERS;
• Dependabot for Maven, Cargo, and GitHub Actions;
• issue forms and pull-request template;
• code of conduct and stronger security/contribution policies;
• package metadata for Maven and Cargo;
• pinned GitHub Actions revisions;
• CI gates for historical PDF integrity, Java verify, Rust fmt, Clippy,
  tests, release build, smoke tests, and diff hygiene.

Scientific boundary

The Rust experiment generator currently uses deterministic SplitMix64 while the
Java benchmark used Java SplittableRandom. These are separately reproducible
measurements, not yet a direct Java-vs-Rust speedup claim.

The next comparison gate is a committed shared experiment plan containing
identical target states and identical Hellman/DP start sets.