Descriptions:
Zubair Trabzada demonstrates how to build a multi-agent AI marketing audit system using Claude Code — a toolkit he frames as replicating work that marketing agencies typically charge $5,000 to $10,000 per month to deliver. The system is structured as 15 Claude Code “skills,” each stored as a skill.md instruction file in a free public GitHub repository.
The flagship command, /market-audit, launches five parallel Claude Code sub-agents that simultaneously analyze a target website across content and messaging, conversion optimization, SEO, competitive positioning, and technical performance. The demo runs the full audit on Calendly.com, producing a scored report (75 out of 100) with an executive summary, category breakdowns, and prioritized findings — exported as a formatted PDF. The video explains how each skill.md file provides Claude with domain-specific instructions, effectively turning a general-purpose assistant into a focused marketing analyst for the duration of the task.
Installation requires Visual Studio Code, the Claude Code extension, and a single terminal command that clones the repository. Trabzada covers the full setup step by step, including how to navigate Claude Code’s permission prompts during agent execution. The parallel sub-agent architecture is the technical centerpiece — each of the five agents operates independently and concurrently, which the video argues is what makes the audit fast enough to be practically useful.
📺 Source: Zubair Trabzada | AI Workshop · Published March 03, 2026
🏷️ Format: Hands On Build
