The Dr. Rob Activator Rule turns every drop in your "Ready to Test for Lucy" folder into a live, gamified, conversion-wired, fully-tagged FFH module — in 15 automated phases, with one human decision point. This page walks through how it works, where Dr. Rob's side plugs in, and how to add your two fellows without losing quality control.
Four lanes. One direction. From Dr. Rob's keyboard to a live URL anyone in the world can visit.
Rob's Claude packages a drop into a standardized handoff envelope and posts it to "Ready to Test for Lucy."
Lucy's Claude runs all 14 phases automatically, pausing only at the Greenlight Gate for one approval.
The cleaned, branded, gamified module + landing page lands here, plus a one-shot deploy script.
GitHub Desktop pushes to main, Vercel deploys to production in about a minute.
Phases tagged DELEGATE call into existing FFH skills. Phases tagged GATE are decision points (one of them is yours, Lucy). Phase 13 (Standardize & Tag) is the canonical-categorization step — every module gets tagged against UN SDGs, education standards, conditions, body systems, FFH gears, and 10 more dimensions so the Network's data layer stays clean. Everything else runs automatically.
Scan the drop folder, build an inventory, read Rob's README if present, classify primary asset type.
AutoOne-page "what is this drop?" summary with proposed slug, coach assignment, and one-line pitch.
AutoValue-add + duplicity check against existing Deploy Ready content. Markdown table for Lucy.
DelegateLucy approves (or auto-approve for simple drops). The only human decision point.
Lucy GateCreate /Deploy Ready/<slug>/ with module + landing + assets + lesson-plan folder.
AutoVoice scrub (Rob keeps mic on health), compliance gates preserved, link cleanup, tracker unification.
DelegateCanonical nav bar + footer, FFH colors, Segoe UI, typography balance, /brand-assets/ paths.
DelegateReal LEARN/LIVE/SHARE It coin assets, GamiPress hooks, rewards pop-ups at ≥3 milestones.
AutoWire to Supabase phit_scores, chna_priorities, ship_priorities + "Powered by PHIT" hub panel.
AutoSMART Clipboard widget + pre-loaded coach greeting (Dr. Rob, Coach Lucy, or co-hosted).
Delegate9-section marketing front door: hero, value prop, 4 audience CTAs, what's inside, sample experience, coach voice, gamification preview, conversion, footer.
AutoSign In / Sign Up / Learn More CTAs on every public page + role-gate decisions for Member-tier content.
AutoCanonical categorization across 15 dimensions: education standards (ISTE, NHES, NGSS), UN SDGs, conditions, body systems, FFH gears, audience tier, coach, coins, careers, geo markets, PHIT priorities, module type, WCAG, compliance, status. Writes meta tags + manifest + Network registry.
NEW · AutoPre-deploy checklist (refuses partial), then hands off to the deploy-ready skill for the Python script.
DelegateAuto-update Airtable Deploy Ledger (with the Phase 13 tags) + Tech Command Center registry + write project memory to CLAUDE.md.
Auto
Every module that ships gets categorized against ALL 15 dimensions — "not
applicable" is a valid answer for some, but every dimension is considered.
Source of truth lives in
FFH Skills/ffh-dr-rob-activator/references/canonical-taxonomies.md
— never hand-invent a value.
<meta> head,
a per-page _tags.json manifest, and the Network-wide
_taxonomy/tag-registry.json. That registry powers SDG-impact
reports to funders, education-standards-coverage decks for school district
sales, conditions-by-module heat maps for clinical partners, and career-pathway
audits for workforce-development grants. Skip tagging once and every downstream
report has holes.
The Activator works today against any drop Rob makes — but every minute Rob's Claude spends standardizing the handoff saves Lucy roughly 10 minutes of cleanse work on the receiving end. A small skill on Rob's side closes that gap.
One skill file (~6 KB), one install command in his Claude project, and a
5-minute walkthrough. He keeps building the way he likes — the envelope skill
just packages his output in the shape Lucy's Activator wants. We'll draft
ffh-rob-handoff-envelope as the next deliverable and pair it
with a "Dear Rob" install guide.
Four humans, two AIs, one production pipeline. The roles below are how a lean tech team of your size stays fast without dropping quality.
| Role | Person | Owns | Git Access | Tier |
|---|---|---|---|---|
| CEO · Director | Coach Lucy | Brand gate, final merge to main, deploy ledger, memory write-back, Activator Greenlight | Owner | Approver |
| CMO · Clinical | Dr. Rob Gillio | Handoff envelope, clinical content authority, voice rule, compliance gates | Maintainer | Approver |
| Fellow #1 | (To onboard) | Feature branches, PRs to main, content drops, page builds under Lucy's review | Collaborator | Contributor |
| Fellow #2 | (To onboard) | Feature branches, PRs to main, content drops, page builds under Lucy's review | Collaborator | Contributor |
| AI Coach | Coach Lucy AI | In-platform wellness coaching, navigation, community activation | None (runtime only) | AI |
| AI Coach | Dr. Rob AI | In-platform clinical education, evidence-based explanations | None (runtime only) | AI |
You asked the right question: do all changes funnel through you, or do fellows push their own commits? Here are the three real options. We strongly recommend Option C — it gives fellows independence without giving up your quality gate, and Vercel's automatic preview deployments make it almost effortless.
Everything funnels through Lucy. Only Lucy pushes to GitHub.
No risk of off-brand or half-baked work shipping Single point of QA Lucy is the bottleneck — work waits on her availability Doesn't scale past 2-3 contributors Fellows can't see their work live without Lucy's helpEach fellow can push directly to main and trigger Vercel deploys.
Maximum speed No bottleneck No QA gate — anything ships, broken or not Hard to roll back, no review trail Security risk if an account is compromisedFellows push to feature branches. Vercel auto-creates preview URLs. Lucy or Rob merges to main when ready.
Every branch gets its own live preview URL automatically Lucy keeps the QA gate without being the only one who can deploy Easy rollback, full audit trail, parallel work Fellows can share preview URLs for feedback before merge ~30 min learning curve on git branching (GitHub Desktop makes it easy)One-time setup per fellow. Do it once and they're productive forever after. Lucy stays the only person who merges to main, so the gate holds.
Settings → Collaborators → Add. Send the invite to their GitHub email.main — Settings → Branches → Add rule for main: ✅ Require pull request review (1 approver), ✅ Require status checks (Vercel preview must succeed), ✅ Do not allow bypassing.Settings → Git → Production Branch = main, Preview Deployments = "All branches". (Default behavior, free on Hobby tier.)File → Clone Repository → theforce.health. Save to a friendly local path.Current Branch → New Branch, name it feat/yourname-what-youre-building. Example: feat/fellow1-prevention-bingo-cardio.~/theforce.health/public/<slug>/, save, see them in GitHub Desktop's diff view.Commit to feat/... with a clear message, then Publish branch (first time) or Push origin.feat/<name>-<short-desc> (new build), fix/<name>-<short-desc> (bug fix), content/<name>-<short-desc> (content update only)./public/, prefix new top-level modules with ffh-. Example: /public/ffh-time-machine/./brand-assets/ — only via the canonical sync script, never hand-edited in the repo./_shared-assets/ (notepad widget, role gate, supabase client) — affects every page, so review carefully./Deploy Ready/ — Lucy owns the orchestration layer.Once you and Rob have walked this page together, here's the suggested sequence to get the full pipeline humming.
Draft ffh-rob-handoff-envelope +
"Dear Rob" install guide. ~1 hour.
Add as collaborator, walk through GitHub Desktop, do a first throwaway PR to learn the flow. ~45 min.
On Rob's next drop, say "fire the Activator" and watch all 14 phases run. Refine from there.