Interactive Modes

Prodara provides five interactive modes that your AI agent uses to investigate, explain, discuss, design, and onboard. Each mode is triggered via a AI prompt and produces structured context that the agent weaves into its response.

Explore Mode

Explore mode performs a read-only investigation of a topic in the product graph. The agent gathers related nodes, edges, diagnostics, and test results without modifying anything. Use it to understand how a feature connects to the rest of the system.

• Traverses graph edges to find related modules, entities, and workflows
• Includes any diagnostics touching the explored nodes
• Shows spec test coverage for the selected area
• Returns structured JSON the agent uses to explain the topic

Help Mode

Help mode provides contextual guidance based on the current project state. Rather than generic documentation, it looks at what you've built so far — your modules, your diagnostics, your coverage gaps — and gives targeted advice on what to do next.

• Analyzes your current graph for missing pieces
• Suggests the next logical spec additions
• Points to specific documentation sections
• Recommends reviewers and tests to run

Party Mode

Party mode kicks off a multi-perspective discussion about a topic. Each built-in reviewer agent contributes its unique perspective — the security reviewer flags risks, the UX reviewer suggests usability improvements, the architecture reviewer evaluates structural impact, and so on.

Design Mode

Design mode generates a structured design document for a proposed feature or change. It outlines the entities, workflows, surfaces, and governance rules needed, along with impact analysis against the existing graph.

• Produces a module-level design with all node kinds
• Includes dependency analysis against the existing graph
• Flags potential conflicts with existing specs
• Generates a design doc the agent can iterate on

Onboard Mode

Onboard mode creates an interactive walkthrough of the project for new team members. It surveys the module structure, key workflows, governance rules, and current state — then produces a structured guide the agent presents step by step.

• Shows the module hierarchy and entity relationships
• Highlights key workflows and their authorization rules
• Summarizes governance policies and constitution
• Includes links to relevant spec tests and documentation

Triggering Interactive Modes

All interactive modes are invoked via the AI prompt file that is generated when you run prodara init --ai <agent>. The agent reads the prompt, gathers context from the compiler, and presents the results in natural language.

Next Steps

Try customizing reviewers to control which perspectives appear in Party mode, or define custom workflows that combine interactive modes with build phases.