# Agentic Driven (AD) 1.0 > The first AI-native methodology for structured development with AI agents. Agentic Driven (AD) is a universal methodology designed for effective human-AI collaboration. It provides clear structure for any project type: software, content, marketing, events, and more. ## 7 Core Principles 1. **Agentic-Driven**: AI agent leads execution, human supervises and decides strategy 2. **Documentation-First**: Document before implementing, update docs with code 3. **Context-Aware**: Manage AI context explicitly via ad.yaml context_files 4. **Contract-Driven**: Define interfaces in docs/interfaces.md before implementation 5. **Continuous Validation**: Validate at every step, exit criteria gate phase transitions 6. **Git-Traceable**: Everything tracked in git with Conventional Commits and semantic versioning 7. **Standards-First**: Use established standards (Conventional Commits, JSON:API, ISO-8601, semver) ## 6 Development Phases ``` v0.0.x → DEFINE Define problem, objectives, scope v0.1.x → DISCOVER Investigate options, viability v0.2.x → DESIGN Design solution, architecture v0.3.x → SETUP Prepare development and validation infrastructure v0.4.x → BUILD Build/create solution v0.5.x → VALIDATE Verify quality, testing v1.0.0 → RELEASE First stable version ``` ## IDE Configuration Files Pre-configured rules for AI assistants: - Cursor: https://agenticdriven.dev/.cursorrules - Claude Code: https://agenticdriven.dev/claude.md - Windsurf: https://agenticdriven.dev/.windsurfrules - GitHub Copilot: https://agenticdriven.dev/.github/copilot-instructions.md - Aider: https://agenticdriven.dev/.aider-rules.md - Continue: https://agenticdriven.dev/.continue-rules.md ## Quick Install ```bash # Download for any IDE curl -sSL https://agenticdriven.dev/download.sh | bash -s # Examples curl -sSL https://agenticdriven.dev/download.sh | bash -s cursor curl -sSL https://agenticdriven.dev/download.sh | bash -s claude ``` ## Project Configuration (ad.yaml) ```yaml domain: "software" # software | book | marketing | event | product | research | course | game mode: "feature" # feature | project context_files: - "README.md" - "docs/decisions.md" active_features: - path: "docs/active/feature-name" description: "Feature description" status: "in-progress" agents: enabled: false ``` ## Key Concepts - **ad.yaml**: Project configuration file defining domain, mode, features, and agent settings - **context_files**: Files AI agents should read to understand the project - **Exit Criteria**: Requirements to complete before advancing to next phase - **Conventional Commits**: Standardized commit message format (feat:, fix:, docs:, etc.) - **Semantic Versioning**: v0.PHASE.ITERATION during development, v1.0.0+ after release ## Links - Website: https://agenticdriven.dev - GitHub: https://github.com/AgenticDriven/agenticdriven - Full Methodology: https://agenticdriven.dev/rules ## License MIT License - Open source, free to use and modify.