WARNING: This is a work in progress prototype

Dot AI is draft standard for creating and managing context in AI coding projects.

Dot AI provides a tool-agnostic way to create and manage context when coding AI projects. It establishes a standardized structure for project documentation, development workflows, and validation processes, enabling more effective collaboration between developers and AI tools.

The framework includes a structured development workflow that guides projects from initialization through feature development and validation, ensuring consistent quality and comprehensive documentation.

Goals

1. Provide context to accelerate project velocity while maintaining software quality.
2. Solve for cases where AI re-implements basic features like auth or middlware.
3. Solve for cases where AI uses the wrong commands due to outdated documentation (e.g. shadcn-ui vs. shadcn@ or NextJS App Router).
4. Create a self documenting, self correcting systems which clear and testable acceptance and validation criteria.

Workflow Overview

flowchart TD
    A[Step 1: Bootstrap Project] --> B[Step 2: Define Context]
    B --> C[Step 3: Document Feature]
    C --> D[Step 4: Implement Feature]
    D --> E[Step 5: Validate Standards]
    E --> F{Validation Passed?}
    F -->|No| G[Step 7: Revise Feature]
    G --> E
    F -->|Yes| H[Step 6: Validation Report]
    H --> I[Step 8: Compliance Review]
    I --> J[Feature Complete]
    J -.-> C[Start Next Feature]

This workflow ensures that all features are properly documented, implemented according to standards, and validated before being considered complete. Each step includes preflight and post-flight hooks that automate validation against established facts and standards.

How can I use this?

Fill out your project details in the various folders. Drop the .cursorrules/.clinerules prompts into your system project, root dir, or settings. Then Cursor/Cline will refer to these requirements and automatically document your project.

CLI Installation

Show hidden files on OS X

defaults write com.apple.finder AppleShowAllFiles -bool true
killall Finder

git clone git@github.com:andrewhopper/dotai.git
cd dotai
npm install # install the packages
npm install -g . # install the CLI globally

CLI Usage

switch to your project dir
e.g. cd ~/dev/proj1
dotai init # this will allow you interactively configure your project

Directory Structure

The .ai directory serves as a standardized location for AI-related context, documentation, and resources. This structure helps maintain consistent organization across projects and enables AI tools to easily locate and utilize project context.

.ai/
└── docs/
    ├── 0-ai-config/                # AI tool configuration
    │   ├── workflow.md             # Development workflow 
    │   ├── mcp.md                  # Model Context Protocol configuration
    │   ├── .cursorrules            # Cursor AI rules
    │   └── .clinerules             # Cline AI rules
    ├── 1-context/                  # Project context documentation
    │   ├── project_context.md      # Project scope and goals
    │   ├── project_conventions.md  # Project-specific conventions
    │   ├── target-personas/        # Target user documentation
    │   └── standards/              # Applicable standards
    ├── 2-technical-design/         # Technical design documentation
    │   ├── development_workflow/   # Development process documentation
    │   │   └── workflow.md         # Standardized workflow with Mermaid diagram
    │   ├── requirements/           # Project requirements
    │   │   └── security/           # Security requirements
    │   │       └── ssl_requirements.md # SSL implementation requirements
    │   └── features/               # Feature specifications
    │       └── [feature-name]/     # Individual feature documentation
    │           └── specification.md # Feature specification
    ├── 3-development/              # Development documentation
    │   ├── folder-locks.md         # Folder lock documentation
    │   └── tasklog/                # AI assistance documentation
    └── 4-acceptance/               # Acceptance and validation
        └── compliance_reports/     # Compliance and validation reports
            └── security/           # Security compliance reports
                └── 2025-03-13_ssl.md # SSL implementation validation

Research

See the following research documents for more information:

• Coding Workflows - Research on Cursor rules and Cline best practices
• MCP Research - Research on Model Context Protocol implementation and best practices

License

This project is licensed under the MIT License - see the LICENSE file for details.

Roadmap

• Create NPM CLI to configure repos for AI
• Create MCP server to manage and maintain context
• Create NPM CLI that will walk users through configuring their apps
• Create implementation plans for each feature with an impact/risk metrics and then adjust levels of human guidance based on the expected risk associated with a set of proposed changes.

Creator

Andrew Hopper
Twitter/X
LinkedIn