Claude Code puts on a suit and starts a marketing agency

What it does

This is a skill system and collection of Python scripts that bolt onto Claude Code. It adds fifteen marketing commands — from website audits and competitor scans to email sequences and PDF report generation — all triggered inside the terminal. The project is essentially a prompt orchestration layer: skill files tell Claude how to analyze a site, while helper scripts fetch pages and format output.

The interesting bit

The architecture leans on parallel subagents — five modules that simultaneously grade a website across content, conversion, SEO, competitive positioning, and strategy, then roll the results into a weighted 0-100 score. It is a neat exercise in prompt engineering as infrastructure: the real code is mostly glue (analyze_page.py, generate_pdf_report.py) holding together a committee of LLM instructions.

Key highlights

• 15 marketing skills, including audits, copy generation, ad creative, content calendars, and client proposals.
• Five parallel subagents score websites across six weighted dimensions (content, CRO, SEO, positioning, brand, strategy).
• Python scripts handle webpage scraping, competitor scanning, calendar generation, and PDF rendering.
• Includes reusable templates for email sequences, launch checklists, and 30-day social calendars.
• Outputs Markdown or optional PDF reports using the reportlab library.

Caveats

• It is strictly a Claude Code add-on; useless without Anthropic’s terminal agent.
• PDF report generation requires an optional extra dependency.
• The scoring methodology is self-defined and not independently validated in the README.

Verdict

Agency builders and solopreneurs who already live in Claude Code will get the most mileage — it is essentially a pre-packaged marketing workflow. If you do not use Claude Code, this repo is just a collection of prompts and Python utilities with no standalone interface.