Helix Core v3.0.0 scientific evaluation evidence: redocking, scalability, workflow fidelity, and structural-alert concordance

Four content-addressed evidence bundles supporting the Helix Core v3.0.0 software paper: Astex-93 controlled redocking, replicated scalability over a locked 500-compound ChEMBL-derived library, a same-engine comparative workflow study, and structural-alert catalog concordance. Includes raw outputs, manifests, runners, independent validators, reports, provenance, and checksums. Redocking does not establish predictive superiority; the comparative study measures orchestration fidelity rather than an independent software comparison; performance timings are from one Windows host; alert concordance is not toxicity validation.