FLWR Branding Studio Kit

Transform your AI coding assistant into a Senior Brand Strategist.

This repository enables you to deploy an autonomous branding agent capable of generating high-level strategic assets (Golden Circle, Archetypes, Tone of Voice) based on raw client intelligence. It uses a strict "Anti-Hallucination" and "Anti-Slop" protocol to ensure professional-grade results.

Developed by FLWR Branding Studio (@flwr.branding)

🚀 Key Features

• RACE Framework Integration: Deep integration of the Research, Action, Context, Example methodology.
• Context Shielding Protocol: Prevents cross-contamination between clients. The structure is fixed; the content is fluid.
• Anti-Slop Writing Checks: Automated rules to prevent generic AI copywriting (no "delve", "tapestry", or sequential adjectives).
• 🎨 Asset Output: Generates assets in a specific Markdown format.
• Automated Project Setup: One-command initialization for new client projects.

📦 What's Included

• .agent/rules/: The "brain" of the strategist (branding_agent_rules.md).
• .agent/workflows/: The step-by-step logic for executing branding projects.
• .agent/templates/: The core templates (Golden Circle, Persona, Voice, etc.).
• scripts/: Automation tools for project scaffolding.

🌍 Compatibility

• Operating Systems: Windows, macOS, Linux (Universal Node.js support).
• IDEs: Optimized for Cursor, Windsurf, and VS Code, but works in ANY editor (Zed, Sublime, etc.).
• Requirements: Node.js (v14 or higher).

🛠️ Installation

1. Clone the repository:

   git clone https://github.com/vansearch/FLWR-Branding-Studio-Kit-.git
   cd FLWR-Branding-Studio-Kit-
   

2. Install Dependencies:

   npm install
   # or
   yarn
   

3. Agent Configuration:

   • If using Cursor, Windsurf, or similar: Ensure the .agent folder is in your workspace root.
   • If using a custom LLM script: Load .agent/rules/branding_agent_rules.md into the system prompt.

⚡ Workflow: How to Run a Project

Phase 1: Initialization

Start a new project for a client. This creates the directory structure and copies the necessary templates.

# Using NPX (Local)
npx flwr-kit "Client Name"

Creates: clients/Client_Name/ with client_intel/ and strategy_output/ folders.

Phase 2: Intelligence Gathering

Dump all raw data into the client_intel folder of your new project.

• Transcripts: Meeting notes, interview recordings (text).
• Briefings: PDF briefs, emails.
• Competitors: Links or text dumps about competitors per the Strategic Framework.

Phase 3: The Strategist (AI Interaction)

Open your AI assistant (Claude 4.5 Sonnet (or higher) or Opus is highly recommended for superior copywriting and nuance) and point it to the project.

1. Activate Protocol:

   "Start branding project for [Client Name]."

2. Interrogation Mode (Automatic): The agent will read your client_intel and stop you if data is missing (e.g., "I don't see target audience data. Please provide X").

3. Strategy Generation: The agent will process the data through the RACE Framework and fill the templates in strategy_output/brand_assets_md/. It uses the Reality Check loop to verify every claim against your provided documents.

Phase 4: Quality Validtion

Before finalizing, the agent runs the Writing Quality Check:

• Checks for forbidden words (AI cliches).
• Ensures strict capitalization rules.
• Verifies distinctiveness vs. competitors.

📁 Repository Structure

.
├── .agent/                  # AGENT BRAIN
│   ├── rules/               # Personality & strict logic constraints
│   ├── templates/           # The source of truth for deliverables
│   └── workflows/           # Step-by-step execution chains
├── docs/                    # Reference documents (Frameworks, Guides)
├── bin/                     # CLI EXECUTABLES
│   └── cli.js               # CLI entry point (npx flwr-kit)
├── package.json             # Node.js dependencies & bin config
├── README.md                # This manual

🤝 Contributing

See CONTRIBUTING.md for guidelines on how to improve the agent's logic or add new templates.

📄 License

MIT