# AINA Launch-Readiness Council — Plan (2026-06-29)

## Verdict
The org is NOT launch-ready. The evidence base shows 30 open issues across 7 departments with no shared critical path, ambiguous launch scope, and three human founder decisions that are blocking the entire agent org from starting. Here is the honest picture.

The single highest-leverage unblocked action in the org costs Jessica 30 minutes: AIN-144 CEO approval to activate the agent roster. Nothing downstream of agent org activation can be started by chartered agents until it clears. This is the only blocker that requires a human decision and costs zero engineering effort.

Three scope decisions must be made in the next 24–48 hours or every department plan is built on a misaligned brief: (1) Is launch-v1 the text/chat/assessment loop with video and audio explicitly deferred, or is it blocked on media-engine? Every department has already implicitly assumed media is deferred — the company goal doc has not caught up. (2) Is 500 modules a hard day-one launch requirement or a scale target? This single unresolved question is the largest time risk in the org — if 500 is a hard requirement, the launch date is undefined. (3) What is the QA launch bar? QA cannot produce a go/no-go verdict without a defined coverage floor and adversarial-pass threshold. Both questions are CEO-level product risk decisions disguised as QA and Curriculum tasks.

The critical path through the org runs: AIN-144 (CEO approval) → AIN-143 (PR bot watcher, release infrastructure live) → AIN-95 (L1–L5 lesson content, Curriculum) → AIN-123 (lesson wiring, Journey — blocked on Curriculum) → AIN-NEW-JP-3 (E2E smoke test) → M7 go/no-go. The cross-department dependency between AIN-95 (Curriculum) and AIN-123 (Journey) is the most dangerous single schedule risk in the org: it has no joint owner, no shared deadline, and no escalation trigger. A named joint owner and a written dependency contract must exist before Curriculum begins authoring so Journey knows when to expect the handoff.

Three legal/infrastructure prerequisites are human actions that no agent can close: Stripe account provisioning (Ali/Richard), DPA + PostHog data-processing agreement (Ali/Laurie — a legal gate before any paid learner is admitted, not an engineering feature), and Gemini Interactions API quota confirmation (Ali GCP action before the tutor spike is dispatched).

The release infrastructure (AIN-143 PR bot, AIN-101 CF flag deploy) is entirely unstarted and is a systemic multiplier — without it, broken code from any of the six other departments can land silently. AIN-143-A is a 1–2 hour task that should be the first build dispatched after AIN-144 clears.

The Growth department shipped the org's first deliverable this session (AIN-134 persona landing variants, PRs #106/#110). Those PRs should not merge until Finch completes a single canon-voice check against positioning-canon doctrine v3.2 — a bounded 2-hour action that closes the brand-drift risk before the Academy URL goes live.

Launch-v1 resolves to five verifiable gates: (A) Core learner loop E2E verified — Éowyn smoke test pass receipt. (B) Legal prerequisites clear — DPA signed, consent UI live, Stripe provisioned. (C) Release infrastructure live — PR bot active, CF flag deploy pipeline live, rollback runbook published. (D) QA launch bar met — coverage floor and adversarial threshold satisfied per Jessica's definition. (E) Canon-voice landing copy live on staging, Finch-verified. When all five are green, Jessica calls launch. Video, audio, personalization substrate, Vertex embeddings, and mobile harness are Phase 2.

## Milestones (M0→M8)

### M0 — Org Activation — Day 1 (2026-06-30)
- **Gate:** AIN-144 CEO approval signed; AIN-143-A PR bot watcher live; AIN-141 second-verifier receipt schema merged; release_marshal confirmed active. No agent build work proceeds until this gate is green.
- **Includes:** AIN-144 (CEO approval gate), AIN-143-A (PR bot watcher scaffold), AIN-141 (PKM second-verifier receipts schema), AIN-NEW-ORG-1 (Gemini Interactions API quota confirmation — Ali GCP action), AIN-NEW-ORG-2 (DPA + PostHog data-processing agreement — Ali/Laurie legal action), AIN-NEW-ORG-3 (Stripe account provisioning confirmation — Ali/Richard), AIN-NEW-PROD-1 (Launch-v1 scope decision: text/chat/assessment loop vs. media-gated — Jessica decision), AIN-NEW-PROD-2 (500-module vs. minimum viable corpus decision — Jessica 48h decision), AIN-NEW-PROD-3 (QA launch bar definition: coverage floor + adversarial-pass threshold — Jessica decision)

### M1 — Release Infrastructure Live — Day 2 (2026-07-01)
- **Gate:** AIN-143 PR bot watcher fully operational with automated review checks on every PR. AIN-101 Cloudflare flag-controlled deploy pipeline live with documented rollback runbook. AIN-143-B (monitoring + alerting on watcher health) green. No PR from any department is merge-ready without bot receipt.
- **Includes:** AIN-143 (PR bot watcher — both sub-tasks), AIN-101 (Cloudflare release control: deploy != release, flags, preview-proof, rollout health), AIN-NEW-REL-1 (Rollback runbook doc)

### M2 — Content Baseline + Platform Scaffolds — Days 3–7 (2026-07-02 to 2026-07-06)
- **Gate:** AIN-95-A coverage matrix committed (no assumed-complete rows). AIN-138-A DO scaffolding live with 10-concurrent load test passing. AIN-NEW-JP-1 Stripe scaffold confirmed with test-mode key. AIN-135 Finch canon mining running. AIN-137 institutional-memory sweep automated. These are parallel independent tracks — all three must clear before M3 opens.
- **Includes:** AIN-95-A (L1-L5 curriculum audit + coverage matrix), AIN-138-A (Durable Object scaffolding + state persistence), AIN-NEW-JP-1 (Stripe account + CF Worker integration scaffold), AIN-135 (Finch canon mining automated), AIN-137 (institutional memory sweep automated), AIN-94-A (Universal Track composer contract doc), AIN-100 (test-coverage gap audit)

### M3 — Core Learner Loop: Platform + Content — Days 8–16 (2026-07-07 to 2026-07-15)
- **Gate:** L1–L2 lesson content verified (AIN-95-B). Entitlement gate + middleware live (AIN-93-A, AIN-93-B). Arena UI functional under Éowyn six-way break test (AIN-138-B). Gemini Interactions API spike result documented (AIN-133-A — go/no-go). Placement rationale UI live (AIN-124-A). Evaluator agents wired for L1–L2 (AIN-95-D partial). All tasks in this milestone require M2 gate to be green first.
- **Includes:** AIN-95-B (Author missing L1–L2 lessons + rubrics), AIN-93-A (Stripe $349 checkout + entitlement flag), AIN-93-B (Entitlement enforcement in middleware), AIN-138-B (Practice arena UI — editor, run, artifact panel, sim runner), AIN-133-A (Gemini Interactions API spike — quota must be confirmed from M0), AIN-124-A (Placement rationale UI + skip-ahead gate), AIN-105 (Gherkin acceptance criteria coverage for journey lanes), AIN-139 (Image/video generation boundary enforcement)

### M4 — Full L1–L5 Content + Legal Prerequisites — Days 17–28 (2026-07-16 to 2026-07-27)
- **Gate:** L3–L5 lessons + rubrics filed (AIN-95-C). All evaluator agents wired and adversarially tested (AIN-95-D, AIN-95-E). DPA signed and consent UI live (AIN-98, AIN-125 — legal gate from M0 must be complete). Privacy/consent layer active before any paid learner is admitted. Canon-voice check on AIN-134 PRs complete (Finch pass). Composer MVP resolves a test learner profile end-to-end (AIN-94-B).
- **Includes:** AIN-95-C (Author missing L3–L5 lessons + rubrics), AIN-95-D (Wire evaluator agents to rubrics — all five levels), AIN-95-E (Adversarial QA on evaluator scoring), AIN-98 (Telemetry events + privacy/consent — PostHog DPA gate), AIN-125 (Drain external telemetry deletion requests with receipts), AIN-126 (Classify pending DER validation receipt edits), AIN-94-B (Universal Track composer MVP — resolves learner path E2E), AIN-103 (Landing/positioning/launch copy — Finch canon-voice verified), AIN-NEW-JP-4 (Video + audio lesson slot spec — so Journey can build the shell now), AIN-NEW-GROWTH-1 (Finch canon-voice review of AIN-134 PRs #106/#110 before merge)

### M5 — Lesson Wiring + Tutor Integration — Days 22–32 (2026-07-21 to 2026-07-31)
- **Gate:** Every L1–L5 lesson node wired to mission card + tutor path (AIN-123-A — depends on AIN-95-C from M4). Gemini Interactions API promoted to production tutor if spike passed (AIN-133-B — depends on AIN-133-A from M3 + quota confirmation from M0). Composer integrated with Journey platform — lesson cards resolve from composer output (AIN-94-C). Backup verifier designated so Éowyn oversubscription does not block the pipeline (AIN-NEW-JP-2). Parallel: Growth landing copy live on staging.
- **Includes:** AIN-123-A (Wire every lesson node to mission card + tutor path), AIN-133-B (Promote Gemini Interactions API to production tutor — if spike passes), AIN-94-C (Integration test composer with Journey platform), AIN-NEW-JP-2 (Backup verifier designation for Journey tasks), AIN-142 (Éowyn adversarial visual QA — full six-way break)

### M6 — QA Launch Bar + End-to-End Verification — Days 30–38 (2026-07-29 to 2026-08-06)
- **Gate:** QA launch bar (coverage floor + adversarial-pass threshold defined in M0-Jessica-decision) is met. AIN-NEW-JP-3 end-to-end smoke test passes (onboarding → assessment → lesson → arena → payment gate), verified by Jared. AIN-100 coverage gap filled to defined floor. AgentOps receipts pipeline (AIN-141) producing auditable second-verifier receipts for all M3–M5 tasks. MCP-first distribution strategy confirmed (AIN-130 go/no-go decision).
- **Includes:** AIN-NEW-JP-3 (End-to-end learner journey smoke test), AIN-100 (Test-coverage gap fill to defined floor), AIN-141 (PKM second-verifier receipts — pipeline producing live receipts), AIN-130 (MCP-first distribution — classify + decision), AIN-NEW-QA-1 (QA launch bar verification report)

### M7 — Launch Go/No-Go Gate — Day 40 (2026-08-08)
- **Gate:** Five launch gates all green: (A) Core learner loop E2E verified — Éowyn smoke test pass receipt exists. (B) Legal prerequisites clear — DPA signed, consent UI live, Stripe provisioned with idempotent webhook. (C) Release infrastructure live — PR bot active, CF flag deploy pipeline live, rollback runbook published. (D) QA launch bar met — coverage floor + adversarial threshold both satisfied per Jessica's M0 definition. (E) Canon-voice landing copy live on staging URL, Finch-verified. Jessica signs go/no-go. AIN-91 and AIN-92 (video/audio) explicitly deferred to post-launch Phase 2.
- **Includes:** Jessica go/no-go sign-off, Five-gate green board, AIN-91 (video) — explicitly deferred to Phase 2, AIN-92 (audio) — explicitly deferred to Phase 2, AIN-96 (personalization substrate) — explicitly deferred to Phase 2, AIN-97 (Vertex embeddings) — explicitly deferred to Phase 2, AIN-102 (mobile harness) — deferred to Phase 2

### M8 — Phase 2: Media + Personalization — Post-launch (TBD after M7 green)
- **Gate:** Media-engine delivers at minimum one rendered video asset + one TTS audio asset in agreed CDN URL format. Personalization substrate pipeline to runtime (AIN-96) operational. Vertex embeddings E2E (AIN-97) operational. Mobile harness scaffolded (AIN-102). AIN-131 NotebookLM trial result documented.
- **Includes:** AIN-91 (video lessons wired from media-engine), AIN-92 (audio/voice TTS in lessons), AIN-96 (personalization substrate pipeline to runtime), AIN-97 (Vertex embeddings/personalization E2E), AIN-131 (NotebookLM trial for video/audio overviews), AIN-102 (mobile harness scaffolding), AIN-79 (ACAD-112 no-wake packet: Laurie evaluator leniency validation)

## The Issue Train (stage gates)

**Intake** (owner: AgentOps (Finch — validates schema and builder/verifier separation))
  - entry: Issue exists in the evidence base or is newly created by a department head. Title, department, lane, priority, and bounded scope are present.
  - exit: Builder and verifier assigned (different people). Issue mapped to a milestone. Depends-on list confirmed. Issue cannot enter Checkout until builder != verifier is explicit.

**Checkout** (owner: release_marshal (enforces depends-on gate; blocks checkout if upstream is open or decision is unresolved))
  - entry: Issue passes Intake. All upstream depends-on issues are in state Done or explicitly deferred with a documented bypass reason.
  - exit: Builder has pulled the issue, confirmed scope is still accurate against current repo state, and posted a start receipt in the issue thread. Any prerequisite human decisions (Stripe account, quota, DPA) are confirmed as resolved before Checkout is allowed.

**Execution** (owner: Builder (lane agent). release_marshal monitors for branches that are stale > 2 days.)
  - entry: Builder has posted a start receipt. Work is in progress on a named branch.
  - exit: (1) PR opened. (2) AIN-143 PR bot review check passes (no merge before bot receipt). (3) Builder posts a completion note with the PR link. Execution gate: no self-verification — builder cannot be the verifier.

**Verification** (owner: Named verifier per issue. Éowyn for visual/adversarial tasks. Jared for platform smoke tests. Finch for canon-voice checks.)
  - entry: PR is open and bot review check has passed. Verifier is notified via issue thread.
  - exit: (1) Verifier runs acceptance criteria against staging (not local). (2) Verifier posts a PKM second-verifier receipt (AIN-141 schema). (3) Receipt is committed to the issue thread. For adversarial tasks (Éowyn): receipt must include break-attempt log. Verifier cannot be the builder.

**Merge-Gate** (owner: release_marshal (AIN-101 flag-controlled pipeline enforces this))
  - entry: Verification receipt exists. No open review threads from the PR bot.
  - exit: release_marshal confirms: (1) PR is up to date with main, (2) no unexplained changes outside the task's bounded scope, (3) all checks green, (4) second-verifier receipt is committed. Then merge is authorized. release_marshal is the sole merge authority — no self-merge by the builder.

**Done / Landed** (owner: AgentOps (Finch confirms PKM receipt); release_marshal confirms no orphan branches)
  - entry: PR merged to main. Branch deleted. No orphan state.
  - exit: (1) Issue state updated to Done in Linear. (2) PKM memory updated (Finch sweeps this via AIN-135/AIN-137). (3) Handoff note posted if the issue has downstream dependents. (4) Milestone gate re-evaluated — if all issues in a milestone are Done, milestone is declared green.

## Decisions for Ali

1. **Launch-v1 scope: text/chat/assessment loop only vs. media-gated launch**
   - Options: Option A (RECOMMENDED): Launch-v1 is the text/chat/assessment/payment loop. Video (AIN-91) and audio (AIN-92) are explicitly deferred to Phase 2 (M8). Company goal doc is updated to reflect this. Every department plans against the text-only scope immediately. | Option B: Launch is blocked until media-engine delivers at least one video asset and one TTS audio asset. Estimate: adds 4–8 weeks to timeline with no committed delivery date from the media-engine dependency. | Option C: Soft gate — launch the text loop but hold the URL 'in preview' until media lands. Risks: preview framing dilutes first-impression quality; unclear when to flip to 'real launch'.
   - **Recommendation:** Option A. No media-engine deliverable exists and both AIN-91 and AIN-92 are explicitly blocked with no unblocking action in-flight. The text/chat/assessment loop is a complete, testable learner experience. Ship it. Update the company goal doc today so every department plans against the same scope.

2. **500-module Universal Track scope: hard day-one launch requirement vs. scale target**
   - Options: Option A (RECOMMENDED): 500 modules is a post-launch scale target. Launch-v1 requires only that the composer can resolve all authored L1–L5 lessons in the verified corpus (however many that is after AIN-95 completes). AIN-94 scope collapses to 'composer resolves authored lessons end-to-end.' | Option B: 500 modules is a hard launch requirement. This dwarfs every other open issue in raw authoring effort, has no coverage baseline, and no velocity data. It almost certainly makes every other milestone slip regardless of engineering execution speed. | Option C: Commit to a specific smaller number (e.g., 50 modules across all five levels) as a launch milestone, with 500 as the 90-day target.
   - **Recommendation:** Option A. The 500-module figure appears to be a scale vision, not a day-one launch contract. Curriculum needs this in writing within 48 hours to scope AIN-94 and AIN-95 correctly. Without it, curriculum_architect cannot plan authoring velocity and will default to treating 500 as a hard requirement — which makes the launch date effectively undefined.

3. **QA launch bar: what coverage floor + adversarial-pass threshold triggers go/no-go**
   - Options: Option A (RECOMMENDED): Define a concrete bar now — e.g., 60% line coverage across critical path files, 78% branch coverage on exercised paths (current baseline), zero adversarial-pass regressions on evaluator scoring, and Éowyn E2E smoke test pass receipt. Adjustable upward after M6 evidence is in. | Option B: Run QA through M6 and let the evidence determine the bar. Risk: QA cannot produce a verdict without a defined threshold — the evidence is meaningless without a pass/fail line. Launch go/no-go becomes subjective. | Option C: Treat the current 24.2% line coverage as acceptable for launch. This is honest about the risk but sets a very low bar for a first paid-learner product.
   - **Recommendation:** Option A. Define the bar now, even if it is adjusted later. The QA department correctly flagged this as a CEO-level decision. Current line coverage (24.2%) against no stated floor is unauditable. A concrete threshold — even a modest one — converts QA output from evidence into a verdict. Jessica should post this number within 24 hours so AIN-100 and AIN-105 have a target.

4. **Gemini Interactions API quota confirmation for AIN-133 (Ali GCP action)**
   - Options: Option A (RECOMMENDED): Ali confirms Vertex/Gemini quota headroom for interactive API usage at expected concurrent learner load (estimate: 50–100 concurrent sessions at launch) before AIN-133-A begins. If quota is insufficient, AIN-133-A spike is deferred and the current tutor stays in production for launch. | Option B: Begin AIN-133-A spike without quota confirmation and treat a 429 response as a go/no-go signal during the spike. Risk: if quota is exhausted in staging, the spike cannot produce meaningful results and the Gemini Interactions API cannot be promoted. | Option C: Defer AIN-133 entirely to Phase 2 and launch with the current tutor implementation.
   - **Recommendation:** Option A. This is an Ali GCP action, not an agent task. It costs 10 minutes to check quota in the GCP console. A tutor that 429s silently at concurrent load is worse for first-impression launch than the current implementation. Confirm or deny before AIN-133-A is dispatched. If quota is not confirmable within 48 hours, default to Option C.

5. **DPA + PostHog data-processing agreement (Ali/Laurie legal action)**
   - Options: Option A (RECOMMENDED): Assign DPA completion to Ali (founder signatory) with Laurie executing the PostHog DPA documentation and consent UI (AIN-98). Set a hard deadline of M4 gate — no paid learner is admitted until the DPA is signed and the consent layer is live. | Option B: Treat DPA as a post-launch legal formality and admit paid learners before it is complete. Risk: legal exposure for processing EU/UK learner data without a compliant agreement. This is not a product tradeoff — it is a legal prerequisite. | Option C: Block launch until DPA is complete regardless of engineering readiness.
   - **Recommendation:** Option A. The DPA is a legal gate, not an engineering task. Separating it from the engineering work (AIN-96/AIN-97 personalization features) and assigning it to Ali + Laurie with a clear M4 deadline is the correct framing. Option B is not a viable business decision for any paying customer. Option C is already embedded in Option A — the M4 gate enforces it.

6. **AIN-134 persona landing PRs (#106/#110): merge now vs. Finch canon-voice check first**
   - Options: Option A (RECOMMENDED): Hold merge of PRs #106/#110 until Finch completes a single canon-voice review pass against positioning-canon doctrine v3.2. This is a 1-agent-pass task, bounded to < 2 hours. Merge immediately after pass receipt. | Option B: Merge now. Accept the risk that landing copy may drift from doctrine v3.2. Revisit copy quality in M4 as part of AIN-103 canon copy work. | Option C: Revert the PRs and re-author from scratch after AIN-103 canon copy work completes.
   - **Recommendation:** Option A. The PRs are already open and close to merge-ready. A Finch canon-voice check is the smallest possible action that closes the brand-drift risk. Option C is unnecessary rework. The org shipped its first Growth deliverable this session — protect it with a fast verification pass, not a rollback.

## New issues to create (13)

- [P0] **AIN-NEW-ORG-1: Confirm Vertex/Gemini Interactions API quota for concurrent learner load (Ali GCP action)** — AgentOps & Factory Health / platform · builder=Ali (founder — GCP console action) · verifier=Richard (CTO — confirms quota is sufficient for 50–100 concurrent sessions) · M0 — Org Activation
- [P0] **AIN-NEW-ORG-2: DPA + PostHog data-processing agreement — legal prerequisite for paid learner admission** — Data, Vertex & Personalization / privacy · builder=Laurie (CDPO — executes PostHog DPA documentation + consent UI AIN-98) · verifier=Ali (founder signatory — signs DPA before any paid learner is admitted) · M4 — Full L1–L5 Content + Legal Prerequisites
- [P0] **AIN-NEW-ORG-3: Stripe account + CF Worker integration scaffold (prerequisite for AIN-93)** — Journey & Platform / platform · builder=Forge (platform builder agent) · verifier=Richard (CTO — account-level confirmation + test-mode key verification) · M2 — Content Baseline + Platform Scaffolds
- [P0] **AIN-NEW-PROD-1: Launch-v1 scope decision — update company goal doc to reflect text-only vs. media-gated** — AgentOps & Factory Health / governance · builder=Jessica (CEO — decision + doc update) · verifier=Monica (Product — confirms department plans align to updated scope) · M0 — Org Activation
- [P0] **AIN-NEW-PROD-2: 500-module vs. minimum viable corpus decision for AIN-94/AIN-95 scope** — Curriculum & Learning Design / curriculum · builder=Jessica (CEO — written decision) · verifier=Curriculum department head (confirms AIN-94/AIN-95 scope is plannable after decision) · M0 — Org Activation
- [P0] **AIN-NEW-PROD-3: QA launch bar definition — coverage floor + adversarial-pass threshold** — QA & Verification / qa · builder=Jessica (CEO — defines the threshold in writing) · verifier=Éowyn (QA — confirms the bar is testable and auditable given current tooling) · M0 — Org Activation
- [P1] **AIN-NEW-JP-2: Backup verifier designation for Journey tasks when Éowyn is over-subscribed** — Journey & Platform / qa · builder=Monica (Product — defines acceptance criteria coverage for backup role) · verifier=Jessica (CEO — approves QA coverage policy) · M5 — Lesson Wiring + Tutor Integration
- [P0] **AIN-NEW-JP-3: End-to-end learner journey smoke test (onboarding → assessment → lesson → arena → payment gate)** — QA & Verification / qa · builder=Éowyn (adversarial QA) · verifier=Jared (platform lead) · M6 — QA Launch Bar + End-to-End Verification
- [P1] **AIN-NEW-JP-4: Video + audio lesson slot spec (CDN URL pattern + metadata shape + fallback behavior)** — Journey & Platform / curriculum-journey · builder=Monica (Product) · verifier=Jared (platform lead) · M4 — Full L1–L5 Content + Legal Prerequisites
- [P1] **AIN-NEW-GROWTH-1: Finch canon-voice review of AIN-134 PRs #106/#110 before merge to main** — Growth, Marketing & Media / growth · builder=Finch (canon/memory agent) · verifier=Head of Growth (confirms copy is usable after Finch review) · M4 — Full L1–5 Content + Legal Prerequisites
- [P1] **AIN-NEW-REL-1: Rollback runbook for CF release pipeline (prerequisite for launch go/no-go gate G-C)** — Release & Platform Control / platform · builder=release_marshal (Releaser) · verifier=Richard (CTO — confirms runbook is executable by a cold agent in < 15 minutes) · M1 — Release Infrastructure Live
- [P1] **AIN-NEW-QA-1: QA launch bar verification report (gate for M6 → M7)** — QA & Verification / qa · builder=qa_automation_engineer · verifier=Éowyn (adversarial QA) · M6 — QA Launch Bar + End-to-End Verification
- [P0] **AIN-NEW-CUR-1: Joint owner + shared deadline for AIN-95 → AIN-123 cross-department dependency** — Journey & Platform / curriculum-journey · builder=Monica (Product — owns the dependency contract) · verifier=Richard (CTO — confirms timeline is realistic) · M2 — Content Baseline + Platform Scaffolds

## Priorities (critical path)

1. AIN-144 — CEO approval gate (Jessica, 30 min, unblocks the entire agent org; nothing downstream starts without it)
2. AIN-NEW-PROD-1 — Launch-v1 scope decision: text loop vs. media-gated (Jessica, 30 min, resolves department planning ambiguity today)
3. AIN-NEW-PROD-2 — 500-module vs. minimum viable corpus decision (Jessica, 48h, gates Curriculum's entire AIN-94/AIN-95 scope)
4. AIN-NEW-PROD-3 — QA launch bar definition: coverage floor + adversarial-pass threshold (Jessica, 24h, without this QA cannot produce a go/no-go verdict)
5. AIN-143-A — PR bot watcher scaffold (release_marshal, 1–2h after AIN-144, makes every other department's output trustworthy)
6. AIN-NEW-ORG-1 — Vertex/Gemini Interactions API quota confirmation (Ali GCP action, 10 min, gates AIN-133-A dispatch)
7. AIN-NEW-ORG-2 — DPA + PostHog data-processing agreement (Ali/Laurie, legal gate, must close before any paid learner is admitted)
8. AIN-NEW-ORG-3 — Stripe account + CF Worker scaffold (Forge, < 4h, gates AIN-93-A pricing work)
9. AIN-NEW-CUR-1 — Joint owner + shared deadline for AIN-95 → AIN-123 cross-department dependency (Monica/Richard, < 1h to draft, prevents the single most dangerous schedule miss)
10. AIN-138-A — Durable Object scaffolding + state persistence (Forge, unblocked, can run in parallel with Stripe scaffold in M2)
11. AIN-95-A — L1–L5 curriculum audit + coverage matrix (curriculum_architect, unblocked, must start before authoring can be scoped)
12. AIN-94-A — Universal Track composer contract doc (curriculum_architect, unblocked, parallel with AIN-95-A)
13. AIN-101 — Cloudflare release control: deploy != release, flags, preview-proof (release_marshal, after AIN-143-A)
14. AIN-135 — Finch canon mining automated (unblocked after AIN-144)
15. AIN-137 — Institutional memory sweep automated (unblocked after AIN-144)
16. AIN-NEW-GROWTH-1 — Finch canon-voice review of PRs #106/#110 (Finch, < 2h, should be a merge prerequisite before those PRs land)
17. AIN-93-A — Stripe $349 checkout + entitlement flag (Forge, after AIN-NEW-ORG-3)
18. AIN-138-B — Practice arena UI (Bridge, after AIN-138-A)
19. AIN-95-B — Author L1–L2 lessons + rubrics (curriculum_architect, after AIN-95-A, parallel with AIN-93-A)
20. AIN-124-A — Placement rationale UI (Mesa, after AIN-93-A)