Note: wizards

Context

The wizards system provides XML-enhanced interactive workflows that guide users through complex, multi-step processes in Attune AI. Each wizard breaks down tasks like debugging, refactoring, or security auditing into structured steps with AI-powered prompts and user interactions.

Architecture

The wizards system separates data types from execution logic:

Core data types (_types.py):

• WizardStep — Defines individual workflow steps with prompts, questions, and execution modes
• WizardConfig — Contains wizard metadata like name, domain, and estimated cost
• WizardResult — Captures completion status, collected data, and execution metrics
• StepType — Enum for step execution modes (question, action, review)

Base execution (base.py):

• BaseWizard — Abstract class that orchestrates step execution, context building, and result processing

Registry management (registry.py):

• Functions to register, retrieve, save, and delete wizard definitions
• Support for both built-in and custom wizards stored as YAML

Built-in wizards

Five specialized wizards extend BaseWizard:

• DebugWizard — Guides systematic debugging workflows
• RefactorWizard — Structures code refactoring processes
• ReleasePrepWizard — Orchestrates release preparation tasks
• SecurityWizard — Facilitates security audit procedures
• TestGenWizard — Assists with test generation workflows

Each wizard customizes build_prompt_context() and process_step_result() to handle domain-specific logic while following the common execution pattern.

Source files

• src/attune/wizards/**

Tags: wizards, interactive