Fullstack MKT Skills: AI Marketing Agent Skill Framework for the Vietnamese Market

Fullstack MKT Skills is an AI marketing agent skill framework specifically designed for the Vietnamese market. It includes 16 professional skills, 4 agent roles, and 3 workflows. By using structured Markdown files, it guides AI tools like Claude Code and ChatGPT to generate professional-level marketing plans, addressing the pain points of generic AI responses that are vague, lack data support, and lack structure.

In AI-assisted marketing practices, generic AI often produces empty content without data, KPIs, or time planning (e.g., a beauty salon marketing plan with only 500 words of vague suggestions). To address this pain point, this project creates a framework exclusive to the Vietnamese market, transforming AI from a chatbot into a professional marketing consultant compatible with mainstream AI tools.

Core Philosophy: "Framework is superior to agent capability". It guides AI to output content according to professional methodologies through structured prompt templates. Core Design:

• 16 skills: divided into four categories—strategic planning (marketing plan, campaign brief, etc.), content creation (video script, ad copy, etc.), data analysis (performance evaluation, marketing report, etc.), and professional fields (SEO, email marketing, etc.);
• 4 roles: strategist (macro planning), content officer (content management), optimizer (data analysis), project manager (campaign execution);
• 3 workflows: monthly planning cycle, campaign execution process, review and improvement cycle; Localization Adaptation: tailored for Vietnamese platforms (Facebook/TikTok/Zalo), payment habits, cultural festivals, regulatory compliance, and language optimization; Usage Method: clone the repository → run the installation script → upload skill files to AI tools → generate content via dialogue, compatible with mainstream AI like Claude Code.

The project showcases multiple practical scenarios:

1. Complete marketing plan: input a monthly budget of 30 million VND for a beauty salon, output a 2000+ word plan including market analysis, KPIs, and timeline;
2. Short video script: input the requirement for a TikTok script of a new lipstick, output two A/B versions of the script with viral score;
3. Performance optimization: input data of CPM 45,000 VND and ROAS 1.8x, output diagnosis and a 48-hour action plan;
4. Reverse budget calculation: input a monthly revenue target of 200 million VND, reverse-calculate the budget and conduct 3 scenario analyses.

This framework proves that prompt engineering can improve the quality of AI's professional output without relying on expensive models. For Vietnamese SMEs, it lowers the threshold for professional marketing consulting; for the AI community, it provides a reference model for encoding domain knowledge into structured prompts, promoting the development of AI-assisted professional work.

The project uses the MIT open source license and encourages community contributions: customize industry-specific skills, create localized versions for other markets, add new skills/workflows, and integrate with automation platforms. Hosted on GitHub, it accepts Issue submissions and Pull Requests for continuous iteration and improvement.