Finish attribution, content QA, and discovery readiness

[advaitmb]

The remaining content and QA issues are addressed: all 66 approved papers now have explicit profile author links, and publication screening uses reviewed inclusive membership years while preserving gaps, omitted ranges, and past decisions.

• Restore the original hero description and append the maintainer-requested current-work sentence; broader rewrites await Tim’s approval.
• Use the approved “Past and present support” wording, retain “Allen School,” and redirect empty legacy sections to useful pages without replacing populated future sections.
• Fix clipped sidebar content in short desktop windows and make the skip-link destination focusable. Verify filters, all 22 profile paper counts, accessibility, reflow, and legacy URLs.
• Add Rongwu’s evidence-backed discovery identity, covering nine current members. An isolated live OpenAlex pilot retrieved 402 unique works and repeated with an identical queue; no production content or review decisions were changed by the pilot.
• Replace stale pre-launch documentation with current issue status, maintainer ownership, QA evidence, and operating instructions.

Validation: npm run check passes (30 JavaScript tests, 41 Python tests, and 106 exported pages); future-content fixture builds pass. Browser checks covered nine routes at five viewport sizes, with zero automated accessibility violations and no horizontal overflow. The large review-queue diff only refreshes membership screening metadata: all 390 observation statuses, targets, decision receipts, and deferrals are preserved.

An isolated clean-checkout update/revert rehearsal also passed without touching production. The maintainer explicitly deferred the IdioFid live signup test. Scheduled maintenance remains disabled; #28 retains reviewer/alert setup, independent monitoring, and two actual weekly cycles. #29 retains the backup-maintainer handoff. Production deployment remains manual.

Closes #22
Closes #25
Closes #26
Closes #27