Verdent Docs

Plan-Code-Verify Workflow

The Plan-Code-Verify workflow provides precise control over requirements, implementation, and validation through multiple input modalities.

What You'll Learn

  • What is Plan-Code-Verify workflow
  • Multi-modal specification methods (natural language, visual, interactive)

Plan-Code-Verify Workflow

Integrates all three phases of Verdent's workflow:

Plan: Clarify requirements and design solutions with clear, visual logic diagrams and plan rules. Express specifications through natural language, UI mockups, or interactive demos.

Code:Review, refactor, and improve your code with ease, and deliver clear results every time. Multi-modal inputs ensure implementation matches exact requirements.

Verify: Catch issues early with a verifier subagent and browser tools for reliability results.

How Plan-Code-Verify Workflow works:

Plan Rules

  • Choose built-in role templates (Frontend, Test, Product Manager)
  • Creating custom plan rules that match your thinking style and work habits

Plan Preview

  • Clarify the structure and content of the plan through plan review
  • After aligning with the AI during the plan review, click ‘Generate Plan’ to create the detailed plan content

Plan Design

Detailed specification including:

  • System architecture diagrams
  • Component structure and relationships
  • Data models and API contracts
  • UI/UX specifications with visual mockups
  • Technology stack justification

Task Breakdown

Hierarchical list with dependencies and priorities to organize implementation workflow

  • Display the real-time progress of tasks
  • Summarize their completion status once finished

Diff Summaries

Review Code Changes and Code-reviewer Subagent

  • Change Localization - Lists all modified files with Code Changes
  • Logic Summarization - Explains what changed and why in natural language
  • Impact Analysis - Highlights downstream effects on documentation, APIs, and dependent code

Verify Earlier

  • Configurable automated verification
  • Verify with Subagents and browser tools

Multi-Modal Specification

Supports multiple ways to express your requirements:

Express requirements, constraints, and acceptance criteria in natural language:

  • Rich contextual understanding of technical and business requirements
  • Supports complex, multi-layered specifications
  • Natural conversation flow for requirement definition

Upload UI mockups, design screenshots, or reference implementations:

  • Pixel-perfect recreation from visual specifications
  • Supports multiple image formats for design communication
  • Visual context for UI/UX implementation

Share working examples or prototypes:

  • Communicate behavior and user experience expectations
  • Demonstrate complex interactions and workflows
  • Live examples for clearer requirement communication

This multi-modal approach ensures complex requirements, especially UI/UX specifications, are captured accurately without ambiguity.


See Also