rootnode-academy/docs/PHASE10_LAUNCH_READINESS.md

Phase 10 — Launch Readiness

Gates 2, 10-14, 19-20, and the final launch matrix. See docs/PHASE10_CANONICAL_STATE_AUDIT.md (Gate 1), docs/SECRETS_AND_INFISICAL.md (Gate 3), docs/PHASE10_PRODUCTION_DEPLOYMENT.md (Gates 4-9/15-16), and docs/PHASE10_SECURITY_AUDIT.md (Gate 18) for the rest.

Gate 2 — Full application audit

Reviewed for architectural coherence across every phase; nothing was found inconsistent or in need of a fix. Identity/auth, learning, Lab platform, business (offerings/registrations/ guardians/agreements), and commerce (pricing/payments/webhooks) all remain exactly as documented in their own phase docs — re-verified by actually rerunning every suite (Gate 19 below), not by re-reading code and assuming it still works.

Gate 10 — Website integration

Update (post-Phase 10.1): the website side is now actually built, not just documented. ced4568/RootNode-academy-website (a separate GitHub repository — Chase provided access and explicitly authorized this phase to build against it) now has a real registration/agreement/ checkout/status flow implementing docs/WEBSITE_REGISTRATION_AND_CHECKOUT.md's full contract: public offering lookup, registration submission (adult and guardian+minor), agreement acceptance, checkout creation, and a status-polling success page that never trusts the redirect URL alone, exactly as that document specifies. Verified there: a clean production build, tsc, eslint, and 9/9 tests all pass. The site's own stale pricing was also corrected to match docs/ROOTNODE_PRICING_AND_OFFERS_V1.md.

It has nothing to talk to yet — ROOTNODE_API_URL/ROOTNODE_INTEGRATION_KEY are unset on the website side (no real Platform API exists to point them at), so every RootNode-facing route on the website returns a clean 503, never a 500, until Gate 4's hardware blocker resolves and a real Platform API origin exists. docs/WEBSITE_REGISTRATION_AND_CHECKOUT.md and docs/ WEBSITE_ENROLLMENT_INTEGRATION.md remain the accurate contract both sides now implement.

Gate 11 — Winter Break offering

Pricing is locked and real: $399.00 launch ("Founding Cohort") / $449.00 standard reference, 39900/44900 cents, paid in full, no deposit, no payment plan — given directly by Chase, validated against real Stripe TEST mode at the exact $399 amount in Phase 9.1. Dates, venue, and capacity remain undecided — not invented by this phase, exactly as every prior phase's own documentation already states. The offering therefore cannot be transitioned to enrollment_open for real, public registration yet — the platform mechanism is ready (PATCH /api/admin/offerings/:id accepts all four fields the moment they're decided).

Gate 12 — Agreements / legal

agreement_definitions/agreement_acceptances architecture is real, tested, and versioned (Phase 7-9). No attorney-approved legal text exists for any required Winter Break agreement — agreement_definitions.body_summary remains a plain-language placeholder, exactly as every prior phase's docs have stated since Phase 7. This phase did not write legal language and did not treat any placeholder as attorney-approved, per Phase 10's own explicit instruction.

LAUNCH BLOCKED — APPROVED AGREEMENT TEXT REQUIRED. Likely required categories (not authored, just the taxonomy Phase 10 asked to identify): program participation terms, guardian consent (minor registrants), a participation/liability waiver, acceptable-use / lab/technology rules, a cancellation/refund acknowledgment tied to whatever policy Gate 14 resolves, and a privacy/data-handling notice if RootNode's counsel determines one is needed beyond what's already in docs/STUDENT_GUARDIAN_DATA_MODEL.md. A media release is optional — RootNode's own choice, not decided here.

Gate 13 — Tax

TAX DECISION REQUIRED. No deliberate Mississippi (or any jurisdiction's) tax treatment has been configured anywhere in this codebase — payment_transactions.amount is charged exactly as the offering's own price, with no tax line item, no tax calculation, and no tax remittance logic anywhere (Phase 9's own explicit "no complex tax automation" scope boundary, unchanged). This is not this phase's decision to make, and the current absence of tax handling should not be read as an implicit "no tax owed" determination — that requires Chase (or a professional) to actually decide, then tell the platform what to do (charge tax as a line item, treat the $399 as tax-inclusive, or confirm no sales tax applies to this specific service in this jurisdiction).

Gate 14 — Refund / cancellation policy

LAUNCH BLOCKED — REFUND/CANCELLATION POLICY REQUIRED. The platform's MECHANICAL behavior is real, tested, and proven against actual Stripe money-movement in Phase 9.1: a refund is recorded as its own tracked state and never automatically cancels the registration or revokes academic enrollment — confirmed directly against the database after a real Stripe TEST refund. What is still missing is the actual customer-facing POLICY text (how many days before the camp is a refund allowed, is it full or partial, etc.) — a founder decision this phase does not invent.

Gate 19 — Complete regression

Fresh stack, this phase, after confirming canonical baseline (Gate 1):

Suite	Result
e2e-phase5b.js	32/32
e2e-phase5c.js	29/29
e2e-phase6f.js	17/17
e2e-phase7.js	17/17
e2e-phase8.js	16/16
e2e-phase9.js	20/20
Course-content validator	PASS, 0 errors
Migration idempotency	Confirmed both ways — a full down -v/up cycle applies all 37 migrations cleanly from empty, and a same-database container restart re-applies with zero errors
Frontend production build	Clean (only the pre-existing chunk-size advisory)
tofu validate (Terraform/OpenTofu)	Clean
ansible-playbook --syntax-check	Clean

131 automated checks, 0 failures. No code changed this phase (Phase 10 is audit/launch-prep, not feature work), so this is confirmation, not new coverage.

platform/test/stripe-integration-test.js (the Phase 9.1 opt-in real-Stripe check) was not rerun this phase — it requires real Stripe TEST credentials, which currently have no durable home (Gate 3's Infisical project doesn't exist yet) and were correctly unset from the local .env after Phase 9.1 ended. Phase 9.1's own evidence (docs/PHASE9_1_STRIPE_TEST_VALIDATION.md) stands as the real-Stripe proof; nothing in this phase's own code changes touches that path.

Gate 20 — Pre-launch dress rehearsal

infra/scripts/drills/test-phase9-winter-camp-commerce.js rerun fresh this phase against current main: 13/13 — website registration → guardian → agreement → checkout → webhook → enrollment → activation → classroom → a real disposable Lab task → instructor visibility, plus failed-payment/duplicate-webhook/abandoned-checkout/last-seat-race, all using the deterministic test provider (no real money, as instructed). This is the closest possible complete rehearsal without real production infrastructure or real dates/venue/capacity — it cannot be run against the ACTUAL Winter Break offering because that offering cannot legitimately open registration yet (Gate 11). No fake evidence was created for the steps genuinely blocked by unresolved business decisions.

Final launch matrix

Phase 10.1 adds a Category column distinguishing four kinds of non-complete gate, since they need entirely different people to resolve them: TECHNICAL READY (code/infra is done, nothing left to build), OWNER DECISION REQUIRED (Chase has to decide something, no code involved), EXTERNAL BLOCKER (a third party — Stripe, an attorney, a landlord — has to act or verify something), HARDWARE BLOCKER (specifically the missing production server, since it gates so many other rows).

Gate	Status	Category	Evidence	Owner action	Blocks registration?
1. Canonical repo reconciled	✅ COMPLETE	—	docs/PHASE10_CANONICAL_STATE_AUDIT.md	None	No
2. Second-PC state reconciled	⚠️ Pending verification	OWNER DECISION REQUIRED	Gitea shows no stranded work; PC #2 itself unverified from here; the real Phase 9.5 pricing doc is confirmed to live there, unpushed	Run §1.3's commands on PC #2, then push the real Phase 9.5 document	No (informational)
3. Secrets centralized (Infisical)	🔧 One access grant away	OWNER DECISION REQUIRED	docs/SECRETS_AND_INFISICAL.md — project exists per Chase, this identity lacks access to it	Grant claude-code-migration access to the existing RootNode Academy project	No (dev secrets already safe locally)
4. Production server	⛔ BLOCKED	HARDWARE BLOCKER	docs/PHASE10_PRODUCTION_DEPLOYMENT.md	Purchase/provision hardware	Yes
5. Proxmox / VMs	⛔ BLOCKED	HARDWARE BLOCKER	Same — no host to provision on	Same	Yes
6. Firewall/isolation	✅ COMPLETE (code)	TECHNICAL READY	docs/PHASE10_SECURITY_AUDIT.md	None until real host exists	No (code ready)
7. Production database	⛔ BLOCKED	HARDWARE BLOCKER	Same hardware blocker	Same	Yes
8. DNS	✅ PARTIAL	HARDWARE BLOCKER (API portion)	rootnodeacademy.com live; no Platform API subdomain yet	Decide/create the API subdomain once a server exists	Yes, for the API portion
9. HTTPS	✅ PARTIAL	HARDWARE BLOCKER (API portion)	Website real HTTPS; Platform API HTTPS pending a real host	Same as above	Yes, for the API portion
10. Monitoring	✅ COMPLETE (code)	TECHNICAL READY	/healthz//metrics live in every test run	None until deployed	No
11. Backups	✅ COMPLETE (code), ⛔ unexecuted	HARDWARE BLOCKER (execution)	Ansible backup role exists, never run	Run once a real host exists	Yes (restore-test)
12. Restore test	⛔ BLOCKED	HARDWARE BLOCKER	Never executed — no target	Same	Yes
13. Stripe TEST validation	✅ COMPLETE	—	docs/PHASE9_1_STRIPE_TEST_VALIDATION.md	None	No
14. Stripe LIVE configuration	⛔ BLOCKED	EXTERNAL BLOCKER	Not configured, not attempted	Chase completes Stripe business/identity verification, provides LIVE credentials	Yes
15. Live webhook configuration	⛔ BLOCKED	EXTERNAL + HARDWARE BLOCKER	Depends on #14 and a real HTTPS endpoint (#9)	Same	Yes
16. Google Workspace	✅ COMPLETE (validated in Phase 5C)	TECHNICAL READY	docs/RUNBOOK_GOOGLE_WORKSPACE_EMAIL.md	Deploy alongside the real server	No (already proven)
17. Website integration	✅ COMPLETE, both sides	—	Contract tested end to end; website-side registration/checkout flow now built and verified (see Gate 10 above)	None	No
18. Winter Break price	✅ COMPLETE	—	$399/$449, real, Stripe-validated, confirmed current with Chase (the live website's own pricing is stale — see docs/PHASE10_CANONICAL_STATE_AUDIT.md)	None (website update is separate, not-RootNode-owned repo work)	No
19. Winter Break dates	⚠️ DECISION REQUIRED	OWNER DECISION REQUIRED	Not decided	Chase decides	Yes
20. Venue	⚠️ DECISION REQUIRED	OWNER DECISION REQUIRED	Not decided	Chase decides	Yes
21. Capacity	⚠️ DECISION REQUIRED	OWNER DECISION REQUIRED	Not decided	Chase decides	Yes
22. Agreements (legal text)	⛔ BLOCKED	EXTERNAL BLOCKER	Placeholder only	Attorney-approved text required	Yes
23. Refund/cancellation policy	⚠️ DECISION REQUIRED	OWNER DECISION REQUIRED	Mechanism proven, policy undecided	Chase decides	Yes
24. Tax decision	⚠️ DECISION REQUIRED	OWNER DECISION REQUIRED / EXTERNAL	Not configured	Chase/professional decides	Yes
25. Production offering	⛔ BLOCKED	OWNER DECISION REQUIRED (compound)	Cannot legitimately open — depends on #19-24	Depends on all of the above	Yes
26. Instructor assignment	🔧 MANUAL OWNER ACTION	OWNER DECISION REQUIRED	Mechanism ready (instructor_offerings)	Chase assigns once the real offering exists	Yes
27. Roster process	✅ COMPLETE (mechanism)	TECHNICAL READY	Rehearsed in Gate 20	None	No
28. Student activation	✅ COMPLETE	TECHNICAL READY	Proven real, twice (Phase 9.1 + Gate 20)	None	No
29. Lab capacity	✅ COMPLETE (code)	HARDWARE BLOCKER (real sizing)	Capacity/race logic proven; real server capacity planning needs #4	Plan once real hardware specs are known	No (mechanism), yes (real capacity)
30. Complete dress rehearsal	✅ COMPLETE (deterministic)	TECHNICAL READY	Gate 20, 13/13	Repeat against the real offering once it can legitimately open	No

Bottom line

PHASE 10 PARTIAL / WINTER BREAK LAUNCH BLOCKED

Every gate that is genuinely software/audit work is complete and verified. Every gate that is a real business, legal, tax, or hardware decision remains exactly that — a decision, not something this phase invented on Chase's behalf. See docs/WINTER_BREAK_LAUNCH_CHECKLIST.md for the single consolidated, categorized checklist Chase can work down.