<p align="center">
<a href="https://dojocoding.io">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/assets/banner-dark.svg">
<source media="(prefers-color-scheme: light)" srcset="docs/assets/banner-light.svg">
<img alt="GTM Engineering Command Center by Dojo Coding: Ads, email and SEO in one growth loop" src="docs/assets/banner-light.svg" width="100%">
</picture>
</a>
</p>
# GTM Engineering Command Center
**A Claude Code plugin that plans, creates, deploys and measures your campaigns across ads, email and SEO, for builders who run their own growth.**
**The autonomous growth operating system for Claude Code.**<br>
Multichannel orchestration. Andromeda-native. Neural pre-testing. Self-learning.<br>
One north star: *print cash.*
[](#installation)  [](LICENSE) 
[Get started](#installation) · [Commands](#slash-commands) · [Agents](#agents) · [Contribute](#contributing) · [Report an issue](https://github.com/DojoCodingLabs/GTM-Engineering-Command-Center/issues/new)
---
## What Is This?
The GTM Engineering Command Center transforms Claude Code into a **full-stack autonomous growth engine**. It doesn't just run ads — it owns the entire revenue lifecycle across every channel, powered by April 2026 operator-level knowledge from $100K-$500K/month spenders:
```
┌─────────────────────────────────────────────────────┐
│ AARRR FUNNEL OWNERSHIP │
│ │
│ Acquisition ──→ Activation ──→ Retention │
│ (Meta, Google, (Onboarding, (Email drips, │
│ TikTok, SEO, Landing pages, Engagement │
│ LinkedIn) Emails) loops) │
│ │
│ Revenue ──→ Referral │
│ (Pricing, (Viral loops, │
│ Upsells, Ambassador │
│ Checkout) programs) │
└─────────────────────────────────────────────────────┘
│
┌─────────┴─────────┐
│ DIAGNOSE → FIX │
│ EMQ audit first. │
│ Then find where │
│ revenue leaks. │
└───────────────────┘
```
**It detects your stack.** Scans your codebase for 20+ integrations (Resend, PostHog, Stripe, Tailwind, Supabase, Vercel, Mux, Sentry, etc.) and adapts.
**It understands Andromeda.** Entity Clustering, ARM, Event Match Quality — the plugin knows how Meta's algorithm actually works in April 2026 and structures campaigns accordingly.
**It diagnoses bottlenecks.** Correlates data across PostHog, Stripe, Meta, Google, email, and SEO to find WHERE your funnel is broken. EMQ audit first — because bad data kills creative.
**It generates assets.** 20+ visually distinct creatives across 4 levers (Persona × Messaging × Hook × Format), landing pages using your design system, email drips matching your template style, SEO content.
**It pre-tests with neuroscience.** Tribe v2 brain prediction scores every creative before deployment. Dead Zone Editing cuts the boring parts. Only winners get budget.
**It verifies its own work.** Playwright, Computer Use, and Chrome DevTools QA every deployment — pixel fires, EMQ checks, mobile responsiveness, funnel walkthroughs.
**It never sleeps.** Cloud Routines pull metrics daily, optimize weekly, and check every PR for conversion tracking completeness.
---
## The Five Laws (from the Media Buying Atlas)
> These laws govern every decision the plugin makes.
| Law | Principle |
|-----|-----------|
| **1. Creative Is the Targeting** | Andromeda's Entity Clustering groups similar ads into one ticket. You need 50-200 visually distinct Entity IDs per week. Minor text variations are dead. |
| **2. Data Hygiene Is the Moat** | ARM reads your conversion data BEFORE your creative. EMQ 8-10 unlocks the full model. You cannot out-creative bad data. |
| **3. Consolidation Beats Fragmentation** | One campaign, 1-2 ad sets, 20-50 diverse creatives inside. The algorithm handles segmentation via Entity IDs and ARM. |
| **4. Agents Run the System, Humans Set Guardrails** | Claude Code agents cut audit time 60%. But unchecked AI optimizes into failure (the Grok Paradox). Human intuition guides; agents execute. |
| **5. Neural Pre-Testing Replaces Vibes** | Tribe v2 predicts brain activation from any creative. Pre-test before spending. The accounts doing this now have the advantage by May. |
---
## The Loop
The `/gtm` command orchestrates the full lifecycle. The `/gtm-quickstart` command walks you through the 7-day implementation playbook.
```
┌──────────────────────────────────────────────────────────────────┐
│ │
│ Phase 1 Phase 2 Phase 3 Phase 4 │
│ ┌─────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │DIAGNOSE │→ │ CHANNEL │→ │ PLAN │→ │ CREATE │ │
│ │EMQ audit│ │ STRATEGY │ │ per │ │ 20+ │ │
│ │funnel │ │ Meta? │ │ channel │ │ distinct │ │
│ │health │ │ Google? │ │ budgets │ │ Entity │ │
│ │score │ │ TikTok? │ │ targets │ │ IDs │ │
│ └─────────┘ │ Email? │ │ copy │ │ 6 hooks │ │
│ │ LinkedIn?│ └──────────┘ └──────────┘ │
│ └──────────┘ │
│ │
│ Phase 4.5 Phase 5 Phase 6 Phase 7 │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ TRIBE V2 │→ │ DEPLOY │→ │ MEASURE │→ │ LEARN │──┐ │
│ │ neural │ │ Meta API │ │ PostHog │ │ save │ │ │
│ │ pre-test │ │ Google │ │ Stripe │ │ insights │ │ │
│ │ dead │ │ Email │ │ Meta │ │ update │ │ │
│ │ zone │ │ TikTok │ │ Google │ │ memory │ │ │
│ │ editing │ │ Git │ │ Email │ │ calibrate│ │ │
│ └──────────┘ │ (PAUSED) │ │ MER │ │ Tribe v2 │ │ │
│ └──────────┘ └──────────┘ └──────────┘ │ │
│ │ │
│ ◄───────────────────────────────────────────────────────┘ │
│ (learnings feed back into next cycle) │
└──────────────────────────────────────────────────────────────────┘
```
---
## Installation
```bash
# Add the marketplace
claude plugins marketplace add DojoCodingLabs/GTM-Engineering-Command-Center
# Install the plugin
claude plugins install gtm-engineering-command-center
# Reload
/reload-plugins
```
Then in your project:
```
/gtm-setup # Detect your stack, configure credentials
/gtm-quickstart # 7-day implementation playbook
/gtm # Full lifecycle orchestrator
```
---
## Slash Commands
### Core Loop
| Command | What It Does |
|---------|-------------|
| `/gtm` | **Full lifecycle orchestrator** — Diagnose → Channel Strategy → Plan → Create → Neural Pre-Test → Deploy → Measure → Learn |
| `/gtm-setup` | First-run wizard — detects 20+ integrations, maps AARRR funnel, creates `.gtm/` config |
| `/gtm-quickstart` | **7-day implementation playbook** — fix data, restructure, create, test, deploy, measure, govern |
| `/gtm-diagnose` | Find the biggest revenue bottleneck. EMQ audit first, then AARRR scoring, Entity ID health check, Blended MER |
| `/gtm-funnel` | Map your AARRR funnel with health scores (0-100) per stage |
### Channel Commands
| Command | What It Does |
|---------|-------------|
| `/gtm-plan` | Multichannel media planning — Meta, Google, TikTok, Email, SEO, LinkedIn, Outreach. Budget-aware structure. |
| `/gtm-create` | Generate 20+ visually distinct creatives across 4 levers and 6 hook categories. Statics, UGC, video, claymation. |
| `/gtm-neurotest` | **Tribe v2 neural pre-test** — brain response prediction, dead zone editing, deploy/test/skip verdicts |
| `/gtm-deploy` | Deploy to any channel — Meta Graph API, **full Google Ads REST `:mutate` route (budget → campaign → adGroup → keywords → RSA, not a stub)**, email provider, git commit. ASC+ and Flood+Underbid support. |
| `/gtm-metrics` | Unified metrics from all channels with Blended MER (total spend / total revenue). **Pulls Google via `google-ads-open-cli` (read-only).** |
| `/gtm-report` | Weekly performance report with cross-channel attribution and AARRR trend |
| `/gtm-learn` | Extract insights, correlate neural scores with actual CPA, calibrate Tribe v2 weights |
### Specialist Commands
| Command | What It Does |
|---------|-------------|
| `/gtm-email` | Create and deploy email drip campaigns using your project's design system |
| `/gtm-seo` | Technical SEO audit + GEO optimization + programmatic content generation |
| `/gtm-landing` | Generate conversion-optimized landing pages using your component library |
| `/gtm-outreach` | Signal-based cold outreach sequences (Tier 1-2 ethical tactics only) |
| `/gtm-referral` | Design and implement a referral program with K-factor projections |
| `/gtm-experiment` | Structured A/B experiment tracking with statistical significance checks |
| `/gtm-qa` | Autonomous QA — pixel verification, EMQ check, mobile testing, funnel walkthroughs |
| `/gtm-catalog` | Create and manage Meta product catalogs; connect to pixels for product-level conversion tracking (uses official `meta ads` CLI) |
### Utility Commands
| Command | What It Does |
|---------|-------------|
| `/gtm-scrape` | Scrape Reddit, X, GitHub for the latest GTM strategies and hacks |
| `/gtm-animate` | Create animated video ads with Remotion |
| `/gtm-routines` | Manage autonomous cloud routines (daily metrics, weekly optimization, PR checks) |
### High-Velocity Advertising (HVA) Commands
The HVA engine — creative discovery at machine speed. A standalone loop parallel to `/gtm` (see **Two Engines** below).
| Command | What It Does |
|---------|-------------|
| `/hva` | **The HVA loop** — fit-check (router) → forge → deploy (PAUSED) → read → decide (cut or compound) → vault |
| `/hva-forge` | Forge a diverse, hypothesis-tagged creative batch (variety over duplication, ≤5/ad set), reusing creative-director + neuro pre-test |
| `/hva-lint` | Lint a campaign structure against the HVA guardrails + Fit Boundary (read-only) |
| `/hva-review` | Read live creatives **by impression volume**, run CLEAR via the deterministic scorer, emit cut/scale/hold per creative |
| `/hva-vault` | Manage the Vault — promote confirmed winners, breed creative families, update Foundry priors, rotate fatigue |
---
## Agents
15 specialized agents. No model pinning — every agent runs on the most capable model in your environment.
| Agent | Role |
|-------|------|
| **funnel-diagnostician** | Cross-channel bottleneck inference engine. EMQ audit, AARRR scoring, Entity ID health, Blended MER. |
| **media-buyer** | Plans campaigns with budget-aware structure (1 ad set at $20/day, ASC+ at $500+/day). Flood+Underbid testing. TikTok + LinkedIn + X.com. Patience Paradox. pLTV bidding. |
| **creative-director** | Generates 20+ visually distinct creatives across 4 levers (Persona × Messaging × Hook × Format) and 6 hook categories. Entity ID diversity. Static priority (60-70% of conversions). AI UGC pipeline. Deliberately absurd formats. |
| **neuro-analyst** | Tribe v2 brain prediction. 7 ROI scores (attention, emotion, memory, decision, comprehension, face response, reward). **Dead Zone Editing** — second-by-second timeline, cut flat signals, front-load peaks. Biological A/B testing. |
| **campaign-operator** | Deploys via Meta Graph API, email providers. **Google Ads REST-`:mutate` deploy (budget → campaign → adGroup → keywords → RSA), brand-exclusion + value-based conversion enforcement.** **ASC+ deployment**, Flood+Underbid, Post ID relaunching, EMQ verification post-deploy. |
| **data-analyst** | Unified metrics. Stripe MRR/LTV/churn, email open rates, SEO rankings, ad performance. **Google Ads metrics via `google-ads-open-cli`.** Cross-channel attribution with Blended MER. |
| **email-marketer** | Designs drip sequences matching your project's existing email design system. Welcome, activation, retention, win-back, upsell. |
| **seo-engineer** | Technical SEO audit, GEO optimization, programmatic content, schema markup. |
| **landing-page-builder** | Generates pages using your project's actual UI library and Tailwind config. Outputs production code. |
| **outreach-operator** | Signal-based cold outreach. Monitors buying signals, generates personalized sequences. Ethical Tier 1-2 only. |
| **referral-architect** | Designs referral programs. K-factor, DB schema + API + UI + email implementation. |
| **stack-detector** | Deep project stack discovery. 20+ integration types → GTM capability mapping. |
| **growth-hacker** | Community intelligence. Reddit, X, GitHub. Compares findings against both Atlas knowledge bases. |
| **qa-engineer** | Autonomous QA via Playwright, Computer Use, Chrome DevTools. Verifies every deployment. |
| **hva-desk** | **The HVA Desk (Layer 3).** Reads insights by impression volume, runs the deterministic CLEAR scorer, converts early algorithm behavior into cut-or-scale calls. Configurable autonomy (recommend / cut-auto / full-auto). Cut fast, scale slow. |
---
## Skills
16 domain knowledge stacks — the brain that agents reference for decisions.
| Skill | Domain | Key April 2026 Knowledge |
|-------|--------|--------------------------|
| `meta-ads` | Meta Ads API + `meta ads` CLI | **Andromeda, Entity Clustering, ARM, EMQ 8+, ASC+, Creative Fatigue Indicator, Attribution Upheaval, Post ID relaunching, native CLI integration (April 2026)** |
| `google-ads` | Google Ads (full parity) | **`google-ads-open-cli` read/measure/audit + GAQL mining → REST `:mutate` deploy. Performance Max, AI Max, Demand Gen, Shopping feed optimization, AI Overviews. 9 rules files + Google Ads Atlas 2026 (18 Parts).** |
| `email-marketing` | Email drips | Deliverability (SPF/DKIM/DMARC), provider APIs (Resend/SendGrid/Postmark) |
| `seo-engineering` | SEO/GEO | Technical SEO, content strategy, schema markup, AI citation optimization |
| `landing-page-patterns` | Conversion pages | Hero patterns, social proof, CTA optimization, design system integration |
| `irresistible-offer` | Copywriting | **6 Hook Categories**, 30-50+ copy variations per visual, angle mining, brand voice locking |
| `campaign-optimization` | Performance | **Patience Paradox**, Creative Fatigue detection, EMQ monitoring, **Blended MER**, budget scaling rules |
| `posthog-analytics` | PostHog API | HogQL, funnel analysis, cohort creation, dashboard templates |
| `funnel-diagnostics` | AARRR bottlenecks | Bottleneck patterns, correlation rules, benchmarks by vertical |
| `stack-detection` | Project scanning | Provider fingerprints (20+ integrations), framework detection, capability mapping |
| `cross-channel-logic` | Multichannel | **TikTok → Meta → Google funnel**, LinkedIn TLAs, X.com, cross-platform creative reuse, budget by business type |
| `stripe-revenue` | Revenue intel | MRR tracking, cohort revenue, LTV calculation, churn diagnostics |
| `browser-qa` | QA verification | Playwright patterns, Computer Use, verification checklists |
| `neuro-testing` | Brain prediction | Brain region mapping, ROI extraction, score interpretation, dead zone editing, setup guide |
| `high-velocity-advertising` | Creative discovery at machine speed | **CLEAR decision rule, the five-layer Stack, the Read Ladder, the two-clocks asymmetry, Benjamini-Hochberg correction, the Vault moat, the Fit Boundary.** The spec behind `/hva*` + `hva-desk` + the deterministic `scripts/hva-score.py` scorer (5 rule files: the-desk, the-foundry, the-vault, guardrails, evidence-base) |
**Plus 3 deep-dive rule files:**
- `andromeda-2026.md` — Full Andromeda mechanics: Entity Clustering, ARM pipeline, EMQ, lifecycle compression, $834M impact study
- `flood-underbid.md` — The contrarian $0 testing method from $80M+ operators
- `ads-cli.md` — Official Meta `meta ads` CLI reference: install, auth via `ACCESS_TOKEN` env var, command catalog, DCO via plural flags, exit codes, fallback list for raw Graph API
---
## Knowledge Base
Three comprehensive knowledge bases totaling 4,500+ lines:
### GTM Creativity Atlas 2026
1,557 lines. 130+ documented tactics. 15 sections. Ethical ratings for every tactic. Complete tools directory with pricing. Case studies from Dropbox, Calendly, Clay, Loom, Notion.
### Media Buying Atlas April 2026
1,337 lines. Compiled from 13 primary research documents, $834M in analyzed ad spend across 3,014 advertisers, and $100K-$500K/month operator playbooks. Covers:
- Andromeda Entity Clustering and ARM mechanics
- The exact campaign structure that $100K-$500K/month accounts run
- ASC+ optimization (62% of ecom spend, 22% ROAS improvement)
- Creative Fatigue Indicator (new April 2026)
- Flood + Underbid contrarian testing ($0 testing at massive scale)
- Tribe v2 Dead Zone Editing workflow
- TikTok, LinkedIn Thought Leader Ads, X.com benchmarks
- Cross-platform budget allocation by business type
- AI UGC pipeline ($28K/day accounts running synthetic avatars)
- Peter Thiel contrarian plays (Patience Paradox, static images > video)
- Complete Whitehat/Grayhat/Blackhat ethical ranking matrix
- 7-day implementation playbook
### AARRR Framework Reference
Industry benchmarks by vertical (developer tools, B2B SaaS, consumer, marketplace). Funnel health scoring algorithm. Cross-stage correlation patterns with diagnostic prescriptions.
---
## Tribe v2 Neural Pre-Testing
The plugin integrates Meta's open-source brain prediction model to score creatives BEFORE deployment:
```
/gtm-create → Generate 20+ visually distinct creatives
│
/gtm-neurotest → Tribe v2 scores each creative:
│
│ Creative Attn Emot Mem Decn Composite Verdict
│ ─────────── ───── ───── ───── ───── ────────── ───────
│ pain_point 82 71 65 78 75.3 DEPLOY
│ social_proof 91 58 72 69 73.8 DEPLOY
│ speed 73 62 48 61 62.1 TEST
│ outcome 67 45 51 55 55.6 SKIP
│
│ Dead zones at 0:04-0:07: CUT. Front-load peak from 0:08.
│
/gtm-deploy → Only DEPLOY + TEST creatives get budget
Budget saved: 40-60% (losers never spend)
```
Three execution modes: Local GPU (fast), Local CPU (slower), Google Colab (free GPU).
After campaigns run, `/gtm-learn` correlates neural scores with actual CPA/CTR and calibrates the composite weights — the system gets smarter every cycle.
---
## Andromeda-Native Campaign Structure
The plugin structures campaigns the way $100K-$500K/month operators run them in April 2026:
| Budget | Structure | Why |
|--------|-----------|-----|
| Under $50/day | 1 campaign, 1 ad set, 15-20 creatives | Can't exit learning with multiple ad sets |
| $50-$100/day | 1 campaign, 1-2 ad sets, 20+ creatives | One broad + one retargeting |
| $100-$500/day | 1 campaign, 2-3 ad sets, 10+ creatives each | One per persona + landing page |
| $500+/day | ASC+ (60-70%) + CBO testing (20-30%) + retargeting (10-15%) | Full Advantage+ Shopping |
**Flood + Underbid** (contrarian testing): 1 CBO, 100+ creatives, low bid cap. Losers spend nothing. Winners self-select. Basically testing for free.
**Patience Paradox**: If CPA ≤ 1.2x target for 5 straight days — LOCK the ad set. No edits for 14 days. Let Andromeda compound.
---
## Two Engines — Slow-Craft and High-Velocity
The plugin ships **two advertising motions.** Match the engine to the conversion, not to the excitement — one company can run both, on different campaigns, in the same account.
| | **Slow-Craft** (`/gtm`) | **High-Velocity** (`/hva`) |
|---|---|---|
| **Run it when** | High-ticket B2B, long sales cycles, weak/offline tracking, low budget, regulated | Fast, cheap, well-tracked conversions: DTC, apps, info products, low-friction lead gen |
| **Reads measured in** | Weeks, by calendar | Hours, by **impression volume** (the Read Ladder) |
| **Quality scored by** | Sales / closed-won in the CRM | Cost per qualifying CAPI event |
| **Creative** | A few high-craft concepts | An inexhaustible, diverse supply (the Foundry) |
| **Decision rule** | Human judgment over time | **CLEAR**, deterministic, at machine speed (the Desk) |
**HVA in one line:** generate, deploy, read, and retire creative hypotheses faster than competitors *without destroying signal quality*, by treating the learning phase as a real-time market maker for creative discovery.
- **The Stack (what to build):** Layer 0 Instrument (clean CAPI) → 1 Foundry (creative) → 2 Structure (funded ad sets) → 3 Desk (CLEAR + Read Ladder) → 4 Vault (the moat).
- **The Loop (what to run):** Forge → Deploy → Read → Decide (cut or compound) → refeed sharper priors.
- **CLEAR:** **C**lean (CAPI truth only) · **L**ead (leading indicators, never first-window ROAS) · **E**conomics (kill at ~3× target CPA, zero conversions) · **A**symmetry (cut fast, scale slow — two clocks) · **R**eplication (Benjamini-Hochberg before crowning a winner).
- **Autonomy is configurable** in `.gtm/config.json` (`hva.autonomy`): `recommend` (default, recommend-only) → `cut-auto` (auto-pause losers) → `full-auto` (auto-pause + bounded auto-scale). Pausing only ever saves money, so it automates a tier before scaling.
- **The decision math is deterministic and unit-tested:** `scripts/hva-score.py` (run `python3 -m unittest discover scripts/tests`). The framework spec lives in the `high-velocity-advertising` skill.
```
/hva spring-tripwire $80/day
→ Fit check: CAPI live, EMQ 9, fast conversion → RUN HVA
→ Forge: 8 distinct concepts (hook × pain × proof × format), neuro-pre-tested
→ Lint + Deploy (PAUSED): one concept per ad, 5 ads in one funded ad set
→ Read (hourly): scorer reads each creative by volume
pain-claymation hard CUT economics-3x-cpa-zero-conv → pause
founder-story hard SCALE confirmed-winner (q=0.03) → scale
ugc-testimonial second HOLD fix-landing-or-offer → fix funnel
→ Vault: promote founder-story; sharper priors refeed the next Forge
```
---
## Routines
Cloud-based autonomous workflows powered by [Claude Code Routines](https://code.claude.com/docs/en/routines).
| Routine | Trigger | What It Does |
|---------|---------|-------------|
| **Daily Metrics Pull** | 9 AM daily | Pulls all channel metrics, saves snapshot, alerts on EMQ drops and CPA spikes |
| **Weekly Optimization** | Monday 10 AM | Full learning loop + funnel diagnosis + weekly report |
| **PR Conversion Check** | GitHub PR event | Reviews landing/email PRs for tracking completeness |
| **Experiment Monitor** | 6 PM daily | Checks A/B experiments for statistical significance |
| **HVA Read Loop** | Hourly | The always-on HVA Desk — pulls per-creative insights, runs CLEAR, acts per `hva.autonomy` (cut fast / scale slow). Hard-blocks if CAPI/EMQ regresses. |
---
## Stack Detection
`/gtm-setup` auto-detects 20+ integration types:
```
Full Stack Detection — 21 integrations detected
| Category | Integration | Details |
|----------------|-------------------|--------------------------------------------|
| Email | Resend | 22 templates, drip scheduler, 3 domains |
| Analytics | PostHog + GA4 | 113+ events, 17 categories, /ph proxy |
| Ads | Meta Pixel + CAPI | Dual tracking, consent-gated, event dedup |
| Payments | Stripe | 3 tiers, 13 Edge Functions, webhooks |
| Design System | Tailwind + Custom | 66+ components, #C980FC/#FF7151/#07070E |
| SEO | Custom seoService | 50+ methods, 9 sitemap routes, JSON-LD |
| Deployment | Vercel | CSP headers, PostHog proxy |
| Video | Mux | Signed playback, uploader, webhooks |
| Error Tracking | Sentry | Frontend + backend, env tagging |
| Feature Flags | PostHog | Dashboard-managed |
| Auth | Supabase Auth | Email + GitHub + Discord + Google + LinkedIn|
| Database | Supabase/PG | 517 migrations, full RLS |
| Real-time | Supabase Realtime | Forum + hackathon live updates |
AARRR Funnel Health:
Acquisition: ████████░░ 75 [STRONG]
Activation: ███████░░░ 70 [STRONG]
Retention: █████░░░░░ 45 [MODERATE]
Revenue: ████████░░ 80 [STRONG]
Referral: ██░░░░░░░░ 15 [WEAK] ← biggest gap
```
---
## The 7-Day Quickstart
`/gtm-quickstart` walks you through the highest-impact implementation:
| Day | Focus | What Happens |
|-----|-------|-------------|
| **Day 1** | Fix Data Layer | CAPI + EMQ audit. Every Purchase event gets email + phone + click ID. |
| **Day 2** | Restructure | Consolidate to ASC+ + CBO + retargeting. Kill the old funnel. |
| **Day 3** | Creative Production | 20-30 distinct Entity IDs. 4 levers × 6 hooks. Statics + UGC + absurd. |
| **Day 4** | Agentic Workflows | Daily metrics, fatigue scan, bleed check, rebalance automation. |
| **Day 5** | Neural Pre-Test | Tribe v2 scores all creatives. Dead zone editing. Deploy winners only. |
| **Day 6** | Cross-Platform | Meta winners → Google Demand Gen + YouTube Shorts + TikTok Spark Ads. |
| **Day 7** | Governance | Blended MER dashboard. Hard rules. Patience Paradox for winners. |
---
## Requirements
| Requirement | Details |
|---|---|
| **Claude Code** | Latest version with plugin support |
| **Meta Ads** | Optional — System User token from live-mode Business app. Plugin auto-installs the official `meta ads` CLI (`pip install meta-ads`) during `/gtm-setup`. |
| **Google Ads** | Optional — Customer ID + developer token. Plugin auto-installs `google-ads-open-cli` (`npm install -g google-ads-open-cli`, read/measure) during `/gtm-setup`; deploys via Google Ads REST `:mutate`. |
| **TikTok Ads** | Optional — Business account for Spark Ads |
| **PostHog** | Optional — Project + Personal API keys for diagnostics |
| **Stripe** | Optional — For revenue diagnostics (MRR, LTV, churn) |
| **Gemini API** | Optional — For Nano Banana 2 creative generation |
| **Tribe v2** | Optional — Python 3.11+ with GPU for neural pre-testing |
| **Email Provider** | Optional — Resend, SendGrid, or Postmark API key |
The plugin works with whatever you have configured. No channel is required — it adapts.
---
## Philosophy
> *A single GTM Engineer with the right agentic stack in 2026 can outperform a 10-person traditional sales and marketing team.*
This plugin is that stack. It's not a tool — it's an autonomous growth team:
- **Media Buyer** who understands Andromeda and structures campaigns the way $100K/mo operators do
- **Creative Director** who generates 20+ visually distinct Entity IDs, not 5 text variations
- **Neuro Analyst** who predicts brain response and edits dead zones before you spend a dollar
- **Data Analyst** who correlates metrics across every channel and calculates Blended MER
- **Growth Hacker** who scrapes the internet for strategies the Atlas hasn't documented yet
- **Email Marketer** who builds drip campaigns matching your design system
- **SEO Engineer** who audits technical SEO and optimizes for AI citations
- **Landing Page Builder** who writes production code using your component library
- **QA Engineer** who verifies everything works in a real browser
- **Funnel Diagnostician** who finds exactly where money is leaking — EMQ first, then AARRR
Human approves. Agent executes. Revenue grows.
---
## Contributing
PRs welcome. This is open source under MIT.
```bash
git clone https://github.com/DojoCodingLabs/GTM-Engineering-Command-Center.git
```
---
## License
MIT — see [LICENSE](./LICENSE). Built by [Dojo Coding](https://dojocoding.io).
<p align="center">
<a href="https://dojocoding.io"><img src="docs/assets/dojocoding-mark.png" alt="Dojo Coding" width="48"></a>
</p>