[ahrazzle]

The site still described the 5 November snapshot and omitted the latest merged research and Virtual Human capabilities. This refresh brings its public narrative through Project Asma #64 and Virtual Human #7, with direct source links and clear distinctions between completed checks, unexecuted plans and open biological questions.

• Adds a four-stage evidence diagram for the CDR2/CDR2L source map and bounded sequence finding: detection, sequence matching, receptor recognition and injury.
• Adds the dated ClinicalTrials.gov snapshot and its overlapping keyword counts, with eligibility and interpretation limits.
• Explains the exploratory frozen-subset plan, updates first-run safeguards and platform checks, and removes stale implementation blockers while retaining execution gates.
• Covers Virtual Human's supported SBML imports, ovarian parameter ledger, research guide and optional evidence bindings. Unchecked SBML units and unpropagated uncertainty remain explicit.
• Consolidates the duplicate Rosen card, replaces the fragile claim total, and records source snapshots, attribution and verification.

Scope is four website/documentation files. Core research methods, datasets, registries and execution rules are unchanged; existing preliminary Pilot values remain withheld. The site uses the newer merged runbook for implementation status where older roadmap text has not caught up.

Validation: all 372 Python tests, repository/link validation, implementation-hash checks, 10,000 artificial slider combinations and offline molecular checks passed. The existing browser suite passed, including GitHub Pages paths and fallbacks. Additional checks covered all expanded notebook sections at 320–1440 px in both themes, no-JavaScript content, unique anchors and no external runtime requests or page errors. A separate Astra review found no material issue. Details and limits: site verification record.

Prepared with OpenAI Codex and GPT-6 Astra agents for coverage, scientific clarity and independent internal review. No new research execution, private-data access or deployment was performed. Internal review and software checks do not establish scientific or clinical validity.