arkaos 2.51.0 → 2.53.0

package/VERSION

@@ -1 +1 @@
-2.51.0
+2.53.0

package/departments/landing/skills/webinar-funnel/SKILL.md

@@ -1,7 +1,8 @@
 ---
 name: landing/webinar-funnel
 description: >
-  Webinar funnel: registration page, reminder sequence, live pitch, replay sequence.
+  Webinar funnel design — registration page, reminder sequence, live pitch
+  script structure, replay/cart sequence, conversion targets.
 allowed-tools: [Read, Write, Edit, Bash, Grep, Glob, Agent, WebFetch, WebSearch]
 ---
 
@@ -21,12 +22,164 @@ does not replace the vault.
 
 # Webinar Funnel — `/landing webinar <topic>`
 
-> **Agent:** Ines (Conversion Strategist) | **Framework:** Webinar Funnel + Russell Brunson
+> **Lead:** Ines (Conversion Strategist) | **Copy:** Teresa (Sales Copywriter) | **Framework:** Russell Brunson Perfect Webinar + False Belief Patterns
 
-## What It Does
+## What ships
 
-Webinar funnel: registration page, reminder sequence, live pitch, replay sequence.
+A production webinar funnel in 6 deliverables:
 
-## Output
+1. **Registration page** with hook + promise + agenda + host + social proof + CTA
+2. **Reminder email sequence** (5 emails, registration → live)
+3. **Live pitch script** with timing markers, false-belief teardown, offer stack
+4. **Replay + cart sequence** (5-7 emails post-webinar)
+5. **Conversion targets** with funnel math (registration → show-up → buy)
+6. **Tracking spec** with pixel events and attribution windows
 
-Webinar funnel map with all pages, emails, and conversion targets
+## The Perfect Webinar Frame (Brunson)
+
+A converting webinar follows a strict structure. Deviating from any element measurably drops conversion.
+
+```
+TIMING               BLOCK            PURPOSE
+─────────────────────────────────────────────────────────────────────
+0:00 - 0:05          Open             Hook, agenda, attendance proof
+0:05 - 0:15          Origin Story     Builder credibility, stakes
+0:15 - 0:45          Content (3 secrets) Teardown of 3 false beliefs
+0:45 - 0:55          Transition       "Here's what to do next"
+0:55 - 1:25          Offer Stack      Itemized value, anchored price
+1:25 - 1:35          Urgency + FAQ    Scarcity reason, top objections
+1:35 - 1:50          Close + Q&A      Repeated CTA, live Q&A
+```
+
+The 90-minute structure is the floor. Live webinars can run 120-180; auto-webinars compress to 60-75.
+
+## False Belief Patterns (the content middle)
+
+The middle of the webinar exists to teardown 3 false beliefs the audience holds — about the **vehicle** (what method), the **internal** (their own capability), and the **external** (their environment / circumstances). Each teardown follows the same shape:
+
+```
+1. State the false belief plainly ("Most people think X")
+2. Tell the origin of that belief (where it came from, why it's plausible)
+3. Tell the story of when YOU believed it (vulnerability anchor)
+4. Reveal the cost of that belief (specific consequence)
+5. Introduce the new belief (what actually works)
+6. Prove the new belief with evidence + story
+```
+
+Pick the 3 most load-bearing false beliefs your offer needs to dismantle. Pick more than 3 and the webinar bloats. Pick fewer and conversion drops.
+
+## Registration Page Anatomy
+
+Above the fold:
+- **Headline** — the hook (curiosity-led specific outcome)
+- **Sub-headline** — the timeframe + named mechanism
+- **Date / time / format** — specific, with timezone
+- **CTA button** — register, not "submit"
+
+Below the fold:
+- **Agenda** — 3-4 bullets teasing the 3 content blocks
+- **Host bio** — one paragraph + photo + 3 credibility markers
+- **Social proof** — 2-3 testimonials or registrant count
+- **Reminder consent** — "We'll send a reminder by email/SMS"
+- **FAQ block** — duration, replay availability, recording disclosure
+
+Conversion target: **30-45%** registration rate from warm traffic, **8-15%** from cold ads.
+
+## Reminder Sequence Template (5 emails)
+
+| # | When | Subject style | Purpose |
+|---|---|---|---|
+| 1 | Immediately | Confirmation + calendar link | Confirm registration, add to calendar |
+| 2 | 24h before | Value-prime story | Story-led pre-sell of the secrets to be revealed |
+| 3 | 1h before | "Starting soon" + login link | Drive show-up |
+| 4 | 10 min before | "We're live in 10" | Last show-up push |
+| 5 | At start | "We're live now" | Catch the last-minute attendees |
+
+Conversion target: **35-55%** show-up rate from registration. Auto-webinar shows higher (60-70%) because re-watching is built in.
+
+## Live Pitch Script Structure (core deliverable)
+
+The script template has fixed sections. Filling each section is the work; structure is not optional.
+
+```markdown
+# Webinar — [TOPIC]
+
+## 0:00 — Open (5 min)
+Hook line: [ONE SENTENCE WITH SPECIFIC NUMBER + TIMEFRAME]
+Agenda: [3 BULLETS — THE 3 SECRETS]
+Attendance proof: [WHO'S HERE / WHERE FROM]
+
+## 0:05 — Origin Story (10 min)
+- The moment everything changed
+- What was at stake (concrete)
+- The cost of NOT solving it
+- The breakthrough insight
+
+## 0:15 — Secret #1 [VEHICLE FALSE BELIEF] (10 min)
+- False belief: [plain sentence]
+- Origin of the belief: [story]
+- When YOU believed it: [vulnerability anchor]
+- Cost of the belief: [specific consequence]
+- New belief: [plain sentence]
+- Proof: [story + evidence]
+
+## 0:25 — Secret #2 [INTERNAL FALSE BELIEF] (10 min)
+(same structure)
+
+## 0:35 — Secret #3 [EXTERNAL FALSE BELIEF] (10 min)
+(same structure)
+
+## 0:45 — Transition (10 min)
+"You're now wondering: how do I actually DO this?"
+- Three options: (1) figure it out alone (slow), (2) hire someone (expensive), (3) follow a proven system (the offer)
+- Introduce the offer name
+
+## 0:55 — Offer Stack (30 min)
+- Component 1: [name + dream outcome + value $X]
+- Component 2: [name + dream outcome + value $X]
+- Component 3: [name + dream outcome + value $X]
+- Bonus 1: [name + scarcity reason]
+- Bonus 2: [name + scarcity reason]
+- Total value: $X
+- Today's price: $Y (anchored vs total)
+
+## 1:25 — Urgency + FAQ (10 min)
+Urgency reason: [scarcity / deadline / cohort cap / price increase]
+Top objections answered: [3-5 with prepared scripts]
+
+## 1:35 — Close + Live Q&A (15 min)
+Repeat CTA + URL
+Live Q&A with prepared questions
+Final close at 1:50
+```
+
+## Replay + Cart Sequence (post-webinar)
+
+| # | When | Subject style | Purpose |
+|---|---|---|---|
+| 1 | 1h after | "Did you miss it?" + replay | Catch no-shows |
+| 2 | 24h | "Quick recap" + replay + CTA | Re-engage, soft pitch |
+| 3 | 48h | Case study | Social proof, momentum |
+| 4 | 72h | FAQ + objection-handling | De-risk |
+| 5 | Cart-close - 24h | "Closing tomorrow" + scarcity | Urgency |
+| 6 | Cart-close - 6h | "Closing in 6h" | Final push |
+| 7 | Cart-close - 1h | "Closing in 1h" | Last-call |
+
+Conversion target: **3-8%** of registrants buy (cold list), **8-20%** (warm list). Replay attribution is 30-50% of total revenue — don't skip it.
+
+## Tracking Spec
+
+Pixel events required:
+- `Registration` — fires on registration form submit
+- `WebinarOpen` — fires on live broadcast page load (50%+ attendance signal)
+- `WebinarComplete` — fires at 75% watched
+- `PitchOpen` — fires when offer URL loads from webinar
+- `Purchase` — fires on order confirmation
+- `ReplayOpen` — fires on replay page load
+- `ReplayPurchase` — fires on order from replay-attributed link
+
+Attribution window: 14 days, last-click weighted toward the live broadcast event.
+
+## Output → Obsidian: `WizardingCode/Webinars/<topic>-<date>/`
+
+Delivers: hook variants + registration page + 5-email reminder sequence + full pitch script + 5-7 email replay/cart sequence + conversion targets + tracking spec.

package/departments/landing/workflows/funnel.yaml

@@ -0,0 +1,135 @@
+id: landing-funnel
+name: Funnel Architecture
+description: Sales funnel architecture matched to audience awareness and price point — squeeze, tripwire, SLO, webinar, VSL, application
+department: landing
+tier: enterprise
+command: "/landing funnel"
+requires_branch: false
+requires_spec: false
+quality_gate_required: true
+
+phases:
+  - id: brief
+    name: Funnel Brief
+    description: Define product, price point, audience awareness level, conversion target
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Frame product, current traffic source, price point, current conversion baseline
+    gate:
+      type: user_approval
+      description: User confirms product, price point, awareness level, conversion target
+
+  - id: awareness-diagnosis
+    name: Awareness Diagnosis
+    description: Diagnose audience awareness level (Schwartz 5 stages) — drives funnel type selection
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Apply Schwartz Awareness Ladder, score audience position
+    gate:
+      type: user_approval
+      description: User approves awareness diagnosis before funnel-type selection
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Funnels/Awareness/"
+        description: Awareness diagnosis with evidence per stage
+
+  - id: funnel-selection
+    name: Funnel Type Selection
+    description: Pick funnel type from Value Ladder matching awareness × price (squeeze / tripwire / SLO / webinar / VSL / application)
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Apply Value Ladder + awareness-price-funnel matrix; select primary funnel
+      - agent_id: saas-strategist-tiago
+        role: Pricing strategy fit check
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves funnel-type selection with rationale
+
+  - id: offer-design
+    name: Offer Design
+    description: Grand Slam Offer construction (Hormozi value equation) for the funnel
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Build offer with dream outcome × perceived likelihood / (time delay × effort)
+    gate:
+      type: user_approval
+      description: User approves offer structure before page architecture
+
+  - id: page-architecture
+    name: Page Architecture
+    description: Map every page in the funnel — squeeze / sales / order / thank-you / upsell / downsell, plus copy framework per page
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Define page list with role, copy framework, conversion target per page
+      - agent_id: sales-copywriter-teresa
+        role: Headline + sub-headline + lead drafts per page
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves the page list and primary CTA per page
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Funnels/Architecture/"
+        description: Page list with copy framework and conversion target per page
+
+  - id: email-sequences
+    name: Email Sequences
+    description: Email sequences supporting the funnel — indoctrination, cart, nurture, abandonment, post-purchase
+    agents:
+      - agent_id: sales-copywriter-teresa
+        role: Email sequence design per supporting role
+    gate:
+      type: auto
+
+  - id: metrics-targets
+    name: Metrics & Targets
+    description: Set conversion targets per page + funnel-level metrics (CAC, AOV, LTV, ROAS)
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Compute conversion baseline + targets, funnel economics
+    gate:
+      type: user_approval
+      description: User approves the funnel metrics targets
+
+  - id: self-critique
+    name: Self-Critique
+    description: Stress-test the funnel — does awareness match offer match price match traffic source?
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Coherence check across all funnel elements
+    gate:
+      type: auto
+
+  - id: quality-gate
+    name: Quality Gate
+    model_override: opus
+    description: Mandatory quality review
+    agents:
+      - agent_id: cqo-marta
+        role: Orchestrate quality review
+      - agent_id: copy-director-eduardo
+        role: Sales copy quality, awareness-match prose, no clichés
+        parallel: true
+      - agent_id: tech-director-francisca
+        role: Page count + metrics target feasibility + tracking integrity
+        parallel: true
+    gate:
+      type: quality_gate
+      required_verdict: APPROVED
+
+  - id: delivery
+    name: Funnel Package Delivery
+    description: Compile the funnel package — pages + copy + emails + metrics targets
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Full funnel package compilation
+    gate:
+      type: auto
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Funnels/"
+        description: Complete funnel package — awareness diagnosis + offer + page architecture + email sequences + metrics targets

package/departments/landing/workflows/webinar.yaml

@@ -0,0 +1,134 @@
+id: landing-webinar
+name: Webinar Funnel
+description: Webinar funnel — registration page, reminder sequence, live pitch script, replay sequence, indoctrination + cart sequences
+department: landing
+tier: enterprise
+command: "/landing webinar"
+requires_branch: false
+requires_spec: false
+quality_gate_required: true
+
+phases:
+  - id: brief
+    name: Webinar Brief
+    description: Define topic, target avatar, live vs evergreen, sale price, dates
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Frame topic, audience, format (live / auto / hybrid), sale price, broadcast dates
+    gate:
+      type: user_approval
+      description: User confirms topic, format, sale price, and dates
+
+  - id: hook-design
+    name: Hook & Promise
+    description: Design the webinar hook (one-time deal, secret framework, mistake list) and the core promise
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Hook variants (3-5), core promise statement, false-belief frames (3)
+    gate:
+      type: user_approval
+      description: User approves hook and core promise
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Webinars/Hooks/"
+        description: Hook variants + core promise + false-belief frames
+
+  - id: registration-page
+    name: Registration Page
+    description: Registration page design — hook, promise, agenda, host bio, social proof, CTA
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Page wireframe with conversion target
+      - agent_id: sales-copywriter-teresa
+        role: Headline + sub-headline + agenda copy + host bio + CTA copy
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves the registration page before email sequence build
+
+  - id: reminder-sequence
+    name: Reminder Sequence
+    description: Email sequence before the webinar — confirmation, value-prime, day-before, hour-before, last-call
+    agents:
+      - agent_id: sales-copywriter-teresa
+        role: 5-email reminder sequence with show-up rate target
+    gate:
+      type: auto
+
+  - id: live-script
+    name: Live Pitch Script
+    description: Webinar script — open, content (3-5 modules), pitch transition, offer, FAQ, close
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Script structure with timing, false-belief teardown, offer stack, urgency, FAQ
+      - agent_id: sales-copywriter-teresa
+        role: Script copy with stories, transitions, CTAs
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves script structure before replay sequence
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Webinars/Scripts/"
+        description: Live pitch script with timing markers
+
+  - id: replay-sequence
+    name: Replay & Cart Sequence
+    description: Post-webinar email sequence — replay availability, encore, cart-open, cart-close, last-chance
+    agents:
+      - agent_id: sales-copywriter-teresa
+        role: 5-7 email post-webinar sequence with cart-close target
+    gate:
+      type: auto
+
+  - id: metrics-targets
+    name: Conversion Targets
+    description: Set registration → show-up → buy targets, plus replay attribution
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Funnel math — registration rate, show-up rate, buy rate, CAC, LTV
+    gate:
+      type: user_approval
+      description: User approves the conversion targets
+
+  - id: self-critique
+    name: Self-Critique
+    description: Stress-test the webinar — does hook match promise match script match offer?
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Coherence check across all webinar elements
+    gate:
+      type: auto
+
+  - id: quality-gate
+    name: Quality Gate
+    model_override: opus
+    description: Mandatory quality review
+    agents:
+      - agent_id: cqo-marta
+        role: Orchestrate quality review
+      - agent_id: copy-director-eduardo
+        role: Script + email copy quality, no clichés, no over-promise
+        parallel: true
+      - agent_id: tech-director-francisca
+        role: Funnel mechanics integrity, tracking, evergreen-vs-live config correctness
+        parallel: true
+    gate:
+      type: quality_gate
+      required_verdict: APPROVED
+
+  - id: delivery
+    name: Webinar Package Delivery
+    description: Compile the full webinar package
+    agents:
+      - agent_id: conversion-strategist-ines
+        role: Full webinar package compilation
+    gate:
+      type: auto
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Webinars/"
+        description: Complete webinar package — hook + registration page + reminder sequence + script + replay sequence + conversion targets

package/departments/marketing/skills/growth-loop/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: mkt/growth-loop
 description: >
-  Design sustainable growth loops (viral, paid, product) that compound over time.
-  Replaces linear funnels with self-reinforcing cycles.
+  Design sustainable growth loops (viral, paid, content, product, community) that compound over time.
+  Replaces linear funnels with self-reinforcing cycles. Spec + math + instrumentation + 30-day experiment plan.
 allowed-tools: [Read, Write, Edit, Agent, WebFetch, WebSearch]
 ---
 
@@ -22,43 +22,160 @@ does not replace the vault.
 
 # Growth Loop Design — `/mkt growth-loop`
 
-> **Agent:** Luna (Marketing Director) | **Framework:** Growth Loops (Andrew Chen / Reforge)
+> **Lead:** Luna (Marketing Director) | **Cross-dept:** Pedro (Paid Specialist) + Tomas (Strategy) + Helena (CFO) | **Framework:** Growth Loops (Andrew Chen / Reforge)
 
 ## Why Loops > Funnels
 
-Funnels are linear: input stops → output stops.
-Loops are circular: output feeds next cycle's input.
+Funnels are linear: input stops → output stops. Every new customer requires fresh top-of-funnel work. CAC stays flat or rises.
 
-## 3 Loop Types
+Loops are circular: **output feeds the next cycle's input**. The system gets stronger with use. CAC trends toward zero on the marginal customer.
+
+A growing business is the sum of working loops, not the sum of funnels.
+
+## Loop Type Decision Matrix
+
+Pick the primary loop based on product DNA, retention shape, and unit economics. Mixing loop types without committing to a primary produces zero compounding.
+
+| Loop type | Default product DNA | Min retention | Min unit econ | Compounding rate | Typical payback |
+|---|---|---|---|---|---|
+| **Viral** | UGC, communication, multiplayer | D30 ≥ 25% | Free or freemium | K-factor (target > 0.5) | Weeks |
+| **Paid** | Clear LTV signal, AOV ≥ CAC × 3 | D30 ≥ 30% | LTV/CAC ≥ 3 | Reinvest ratio | 6-12 months |
+| **Content** | Searchable problem space, repeat searcher | Brand recall | Long-tail compound | Page-month growth | 3-12 months |
+| **Product** | Multi-user by design (invite, share, embed) | D30 ≥ 40% | Free trial / freemium | Natural rate of growth | Months |
+| **Community** | Identity-driven adoption, practitioner audience | D90 ≥ 50% | LTV scales with engagement | Member-second-order growth | Quarters |
+
+If you don't pass any minimum, **fix retention first**. Loops on broken retention amplify churn.
+
+## The 5 Loop Types Fully Spec'd
 
 ### 1. Viral Loop (User-Generated)
+
 ```
-User creates content → Content indexed/shared → New user discovers → Signs up → REPEAT
+User creates content → Content indexed/shared → New user discovers → Signs up → Creates content → REPEAT
 ```
-Example: Pinterest, YouTube, Notion templates
-Key metric: Viral coefficient (K-factor > 1 = exponential)
+
+Worked example: Pinterest, YouTube, Notion templates, Figma community files.
+
+Key metric: **K-factor** = invites sent × conversion rate. K > 1 = exponential. K = 0.5 means each new user brings 0.5 more (sub-viral but compounds with paid).
+
+Failure modes: K declines as audience saturates → loop dies. Invite incentives without product value → users opt out. Content production without index → SEO doesn't pick up.
+
+Instrumentation needed: sign-up event with source (viral_invite vs organic), invite-sent event with recipient_id, conversion event from invite landing.
 
 ### 2. Paid Loop (Revenue-Funded)
+
 ```
 User pays → Revenue reinvested in ads → New user acquired → Pays → REPEAT
 ```
-Example: DTC brands, SaaS with efficient paid
-Requirement: LTV:CAC > 3:1, payback < 12 months
 
-### 3. Product Loop (Built-in)
+Worked example: DTC brands with strong AOV, SaaS with efficient paid (Notion, Webflow at scale), pure-play e-commerce.
+
+Required unit economics: **LTV/CAC ≥ 3, CAC payback < 12 months**. Below those, the loop runs negative against runway.
+
+Failure modes: Channel saturation drives CAC up faster than LTV improves. Single-channel concentration (90% Meta) collapses on platform changes. Attribution opacity prevents real ROI measurement → you reinvest into losing channels.
+
+Instrumentation needed: per-channel CAC tracking, cohort LTV (D30/D90/D365), payback cohort analysis, attribution model (last-touch + multi-touch comparison).
+
+### 3. Content Loop (Search-Driven)
+
+```
+User searches problem → Lands on content → Converts to lead/customer → Content earns links → Ranks higher → More searchers find it → REPEAT
+```
+
+Worked example: Ahrefs, HubSpot, Investopedia, Wirecutter.
+
+Key metric: **Page-month growth rate** + **Domain Authority compounding**. New page-months publish faster than old page-months decay.
+
+Failure modes: Content quality below SERP threshold → no ranking. No internal linking → no topical authority. AI-flood post-2024 raised the quality floor; thin content doesn't rank.
+
+Instrumentation needed: GSC clicks/impressions per page, organic conversion rate per page, time-to-first-rank, total indexable pages.
+
+### 4. Product Loop (Built-in Virality)
+
+```
+User uses product → Product surface exposes to others → Others see it / receive output → Sign up / install → REPEAT
+```
+
+Worked example: Slack (invite team), Dropbox (share file), Calendly (send invite link), Figma (share design link), Loom (share recording).
+
+Key metric: **Natural rate of growth** = % of new users from product surface (not paid, not virally invited).
+
+Failure modes: Product output stays inside the team (no external exposure surface). Sharing requires extra steps (high friction). Output looks bad outside the product (no brand impression).
+
+Instrumentation needed: source attribution on first session (product_share_url vs organic vs paid), share-link click-to-signup conversion, output-publish event tracking.
+
+### 5. Community Loop (Identity-Driven)
+
 ```
-User uses product → Product exposes to others → Others join → REPEAT
+Practitioner joins community → Status from contributing → Other practitioners notice → Join → Contribute → REPEAT
 ```
-Example: Slack (invite team), Dropbox (share files), Calendly (send invite)
-Key metric: Natural rate of growth (organic + viral, no paid)
 
-## Loop Mapping Process
+Worked example: Stack Overflow, Reddit communities, Substack networks, Slack practitioner groups, Discord servers.
+
+Key metric: **Member-second-order growth** — for every new member, how many new members do they bring within 90 days?
+
+Failure modes: Community grows but doesn't tie to product revenue (community-led ≠ revenue loop). Status hierarchy collapses → top contributors leave. Spam erodes signal → quality contributors disengage.
+
+Instrumentation needed: cohort retention curves (D30/D90/D365), contribution distribution (Gini coefficient), second-order growth attribution.
+
+## Loop Specification Format
+
+Every loop spec must answer these 6 questions in the same shape:
+
+```yaml
+loop_name: <descriptive name>
+loop_type: viral | paid | content | product | community
+
+step_1_trigger:
+  what_happens: <observable event>
+  mechanism: <named mechanism, not vibes>
+  metric: <measurable target>
+  owner: <named human>
+
+step_2_action:
+  what_user_does: <observable user action>
+  mechanism: <what makes this action likely>
+  metric: <conversion rate target>
+  owner: <named human>
+
+step_3_output:
+  what_gets_produced: <artifact, signal, content, money>
+  visibility: <internal | shareable | indexed | broadcast>
+  metric: <output volume target>
+  owner: <named human>
+
+step_4_reinput:
+  how_output_becomes_input: <named mechanism>
+  audience: <who sees the output>
+  metric: <reinput conversion rate>
+  owner: <named human>
+
+cycle_time: <hours / days / weeks per full loop>
+compounding_factor: <K-factor / reinvest ratio / page velocity / NRG / 2nd-order rate>
+breakeven_threshold: <the number where loop self-sustains>
+```
+
+## Loop Design Checklist (pre-instrumentation)
+
+Before instrumenting, the loop must pass these 8 checks. Failing any check means the loop is incoherent — fix the spec before tracking.
+
+- [ ] **Output is observable** — the loop's output (artifact, signal, post, share) can be counted, not inferred
+- [ ] **Re-input mechanism is named** — "users tell their friends" is vibes; "invite link with referral code on Day 7 prompt" is a mechanism
+- [ ] **No infinite resource assumed** — the loop doesn't require infinite content production, infinite paid budget, or infinite founder time
+- [ ] **Step ownership is human-named** — every step has a named owner, not "marketing team"
+- [ ] **Cycle time is sub-quarterly** — at least one full loop completes within 90 days, otherwise feedback is too slow
+- [ ] **Compounding factor is computable** — you can express the math: K-factor, reinvest ratio, NRG percentage
+- [ ] **Breakeven threshold is named** — you know the number at which the loop self-sustains
+- [ ] **Failure modes are listed** — the spec names 2-3 things that would kill the loop and the early signals
+
+## Common Failure Modes Across All Loop Types
+
+1. **Mistaking content for loop** — publishing content monthly isn't a loop unless each piece earns links/searches that bring new users
+2. **No instrumentation** — without per-step tracking, you can't tell if the loop is working or if growth is from somewhere else
+3. **Multiple half-loops** — running 3 loops at 0.3x compounding ≠ 1 loop at 1.0x. Pick a primary
+4. **Loop without retention** — a loop on broken retention amplifies churn. Fix D30 first
+5. **Loop ignores unit economics** — viral loops still need someone to pay eventually; product loops still need conversion rate
 
-1. **Identify the trigger** — What starts the cycle?
-2. **Map the user action** — What does the user do?
-3. **Find the output** — What does the action produce?
-4. **Trace the reinput** — How does output bring new users?
-5. **Measure the cycle** — How long is one loop? What's the conversion at each step?
-6. **Optimize the bottleneck** — Which step has the biggest drop-off?
+## Output → Obsidian: `WizardingCode/Marketing/GrowthLoops/<product>-<date>/`
 
-## Output → Growth loop diagram + metrics per step + optimization plan
+Delivers: retention baseline + loop type selection rationale + full loop spec (all 4 steps with mechanism/metric/owner per step) + compounding math + unit economics check + instrumentation tracking spec + 30-day experiment plan + 1-page executive summary.

package/departments/marketing/workflows/growth-loop.yaml

@@ -0,0 +1,141 @@
+id: mkt-growth-loop
+name: Growth Loop Design
+description: Design a sustainable growth loop (viral / paid / content / product / community) that compounds over time
+department: marketing
+tier: enterprise
+command: "/mkt growth-loop"
+requires_branch: false
+requires_spec: false
+quality_gate_required: true
+
+phases:
+  - id: brief
+    name: Growth Loop Brief
+    description: Frame product, current growth, audience, retention baseline, runway
+    agents:
+      - agent_id: marketing-director-luna
+        role: Frame product, current new-user volume, current retention curve (D1/D7/D30), runway
+    gate:
+      type: user_approval
+      description: User confirms product context, current growth, retention baseline
+
+  - id: retention-baseline
+    name: Retention Baseline
+    description: Measure the current retention curve and identify the activation moment
+    agents:
+      - agent_id: marketing-director-luna
+        role: Compute D1/D7/D30/D90 retention from existing analytics, identify activation event
+      - agent_id: strategy-director-tomas
+        role: Compare retention to category norms; flag whether product-market fit is plausible
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves retention baseline and activation event
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Marketing/GrowthLoops/Baseline/"
+        description: Retention baseline with activation event and category comparison
+
+  - id: loop-selection
+    name: Loop Type Selection
+    description: Pick the primary loop type (viral / paid / content / product / community) based on product DNA and retention curve
+    agents:
+      - agent_id: marketing-director-luna
+        role: Apply loop-type fitness matrix (retention shape × CAC payback × content production capacity)
+      - agent_id: paid-specialist-pedro
+        role: Paid-loop feasibility math (CAC × LTV × payback window)
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves loop type + secondary loop choice
+
+  - id: loop-design
+    name: Loop Specification
+    description: Spec the loop — Input → Action → Output → Re-input. Define every step's mechanism and metric
+    agents:
+      - agent_id: marketing-director-luna
+        role: Loop spec with named mechanism per step and target metric per step
+    gate:
+      type: user_approval
+      description: User approves the loop spec before instrumentation
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Marketing/GrowthLoops/Specs/"
+        description: Loop specification with per-step mechanisms and metrics
+
+  - id: compounding-math
+    name: Compounding Math
+    description: Compute the loop's compounding factor — viral coefficient K, content velocity, CAC payback — and the conditions for self-sustaining
+    agents:
+      - agent_id: marketing-director-luna
+        role: Compute K-factor / content compounding rate / payback window; identify the breakeven threshold
+      - agent_id: cfo-helena
+        role: Validate unit economics — LTV/CAC, contribution margin, runway impact
+        parallel: true
+    gate:
+      type: user_approval
+      description: User approves the loop economics
+
+  - id: instrumentation
+    name: Instrumentation
+    description: Define the tracking spec — every loop step needs an event + a dashboard widget
+    agents:
+      - agent_id: marketing-director-luna
+        role: Event schema per step, dashboard layout, owner per metric
+      - agent_id: tech-director-francisca
+        role: Tracking implementation feasibility check
+        parallel: true
+    gate:
+      type: auto
+
+  - id: experiment-plan
+    name: 30-Day Experiment Plan
+    description: First 30 days — what to test, what hypothesis per test, how to declare a win
+    agents:
+      - agent_id: marketing-director-luna
+        role: 3-5 prioritized experiments with hypothesis, MDE, success criteria
+    gate:
+      type: user_approval
+      description: User approves the 30-day experiment plan
+
+  - id: self-critique
+    name: Self-Critique
+    description: Stress-test the loop — is every step instrumented? Are dependencies named? What kills the loop?
+    agents:
+      - agent_id: marketing-director-luna
+        role: Loop coherence + failure-mode enumeration
+    gate:
+      type: auto
+
+  - id: quality-gate
+    name: Quality Gate
+    model_override: opus
+    description: Mandatory quality review
+    agents:
+      - agent_id: cqo-marta
+        role: Orchestrate quality review
+      - agent_id: copy-director-eduardo
+        role: Loop documentation clarity, no clichés, mechanism specificity
+        parallel: true
+      - agent_id: tech-director-francisca
+        role: Instrumentation correctness, tracking event integrity, dashboard feasibility
+        parallel: true
+    gate:
+      type: quality_gate
+      required_verdict: APPROVED
+
+  - id: delivery
+    name: Growth Loop Package Delivery
+    description: Compile the loop package — spec + economics + instrumentation + experiment plan
+    agents:
+      - agent_id: marketing-director-luna
+        role: Full growth loop package + 1-page executive summary
+    gate:
+      type: auto
+    outputs:
+      - type: document
+        format: markdown
+        obsidian_path: "WizardingCode/Marketing/GrowthLoops/"
+        description: Complete growth loop package — retention baseline + loop spec + compounding math + instrumentation spec + 30-day experiment plan + exec summary

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "arkaos",
-  "version": "2.51.0",
+  "version": "2.53.0",
   "description": "The Operating System for AI Agent Teams",
   "type": "module",
   "bin": {

package/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "arkaos-core"
-version = "2.51.0"
+version = "2.53.0"
 description = "Core engine for ArkaOS — The Operating System for AI Agent Teams"
 readme = "README.md"
 license = {text = "MIT"}