Lesson 5.4: Your AI Media Buying Agent

You're about to install a full AI media buying system — 5 specialized agents, 500+ knowledge files, and built-in commands. Download the zip, drop it into your Claude Code workspace, and you've got an entire traffic team that works 24/7.

🎯 What You're Getting

This isn't just one file — it's a full Traffic Command system. 5 specialized AI agents that work together, 500+ knowledge files from real campaigns, and built-in commands you can run. Drop it into your workspace and you've got an entire traffic team.

🎖️

Aurelius

Traffic Commander — routes tasks to the right specialist

📊

Zenith

Metrics Analyst — reads data, diagnoses problems

🎨

Vortex

Creative Architect — writes copy, generates hooks & angles

🔍

Nexus

Tracking Specialist — pixels, CAPI, attribution

🧠

Stratos

Strategy Advisor — scaling, budgets, campaign structure

Plus 500+ knowledge files from real campaigns, 18 skills, 5 built-in commands, and client report templates. This is the same system we use in our agency.

⬇️ Step 1: Download the Agent

📦 Traffic Command — Multi-Agent System

5 agents, 500+ knowledge files, 18 skills, 5 commands. Your entire AI traffic team.

⬇️ Download Traffic Command (.zip)

agents/ — 5 specialized AI agents

knowledge/ — 500+ campaign frameworks

skills/ — 18 plug-and-play skills

commands/ — 5 built-in commands

📂 Step 2: Install It

Unzip the downloaded file — you'll see a folder called traffic-command-student-export
Open VS Code with your my-ad-agency folder (from Lesson 5.2)
Drag the entire contents of the unzipped folder into your workspace
You should see folders: agents/, knowledge/, skills/, commands/ plus a README.md
That's it. Your traffic team is loaded.

⚠️ If you customized your old CLAUDE.md (added client names, your voice preferences, etc.), don't delete it. The agents folder works alongside your existing CLAUDE.md — they add to it, they don't replace it.

🧪 Step 3: Test It

Open Claude Code and try these. If it responds with detailed, specific answers, your agent is working.

TEST 1 — Diagnosis (uses Zenith)

"My CPM is $45 and CTR is 0.6%, what's wrong?"

TEST 2 — Creative Generation (uses Vortex)

"Write me 10 Facebook ad hooks for a dentist in Miami that needs more patients"

TEST 3 — Strategy (uses Stratos)

"My client is a gym owner with $2K/month ad budget. Build me a campaign plan."

✅ Good sign: If it references the knowledge base (TFC files), mentions Andromeda, uses specific niche benchmarks, or routes your question to the right specialist agent, the system is loaded and working.

🧠 What's Inside the Agent

Here's what your agent knows. You don't need to memorize any of this — the AI has it built in. But it helps to know what's under the hood.

📂 500+ Knowledge Files

Real campaign data, frameworks, and strategies from actual ad accounts. Andromeda paradigm, creative diversification, hook frameworks, niche benchmarks, scaling protocols. The agents pull from this knowledge base when answering your questions — so you're getting answers based on real results, not generic AI.

🛠️ 18 Skills

Plug-and-play capabilities: hook generation, copy generation, creative briefs, metric diagnosis, creative fatigue detection, funnel analysis, tracking audits, campaign structure, budget allocation, kill/scale rules, unit economics, and more. Each skill is a focused workflow the agents use automatically.

⚡ 5 Built-in Commands

Type these in Claude Code for instant workflows: /traffic (general traffic help), /traffic-audit (full campaign audit), /traffic-diagnose (diagnose a problem), /traffic-hooks (generate hooks), /traffic-scale (scaling plan).

📋 Templates

Ready-to-use templates for performance reports, creative briefs, campaign audits, scaling readiness checklists, and campaign launch checklists. Just ask and the agents fill them in with your actual data.

🔍 Diagnostic Framework

Zenith diagnoses problems automatically: High CPM + low CTR = creative problem (routes to Vortex). Good CTR + low conversions = landing page problem. Good apps + low shows = follow-up problem. It tells you what's broken, whose job it is to fix it, and what to do next.

✅ Setup Checklist

Installation Progress 0 / 0 complete

Downloaded the agent zip file Click the blue download button above
Unzipped and see agents/, knowledge/, skills/, commands/ folders Plus README.md and DEVELOPER-GUIDE.md
Dragged all contents into your VS Code workspace Your workspace should now have the agents, knowledge, skills, and commands folders
Ran Test 1 — asked for ad hooks Should give you hooks using the 7 categories
Ran Test 2 — asked about a high CPL Should tell you to kill it (3x+ target for 4 days)
Ran Test 3 — asked for a campaign plan Should mention Advantage+, creative diversity, and niche benchmarks

🚀 What's Next

Your agent is installed. Now you're going to learn how to actually use it.

➡️ Lesson 5.5: Reading Your Ad Data With AI — pull your campaign data into Claude Code and let the agent tell you what's working and what's not.

💡 Pro tip: As you go through the next lessons, your agent already knows everything they teach. So if you ever forget a concept, just ask your AI. That's the whole point.

🏷️ Tags

agentCLAUDE.mddownloadmedia-buyingandromedahooksoptimizationniche-benchmarks

← Meta API Connection All Lessons Reading Ad Data →