Custom Prompts for AI Coding Agents

This project provides a set of prompts for AI coding agents to implement a structured, two-step Plan & Execute workflow, inspired by the principles of Kiro.

The primary motivation behind this methodology is to minimize the time and prompt iterations required to generate working code, improve overall code quality, and enable seamless switching between AI coding agents within the same project.

Inspired by Kiro-style Prompts for Spec-Driven Development, this implementation significantly simplifies documentation-driven workflow. Feature specifications are captured in a dedicated prd.md file, creating a shared source of truth that persists beyond individual chat sessions. This enables team collaboration and work resumption while utilizing Planner and Executor operational modes.

The Core Workflow: Plan & Execute

The methodology is split into two distinct phases, each handled by a dedicated custom AI agent mode.

1. Planning phase (Planner mode): The AI acts as a senior architect. You provide a high-level feature description, and the AI guides you through an interactive process to create a complete technical specification.
2. Execution phase (Executor mode): The AI acts as a meticulous engineer. It reads the approved specification and implements the feature, ensuring strict adherence to the plan.

Setting Up the Workflow

This workflow uses two distinct AI personas, each with a specific role. The setup process varies between different AI coding tools.

Cursor (custom modes)

1. Planner mode

• Name: "Planner"
• Model: Gemini 2.5 Pro or Claude Opus 4 (powerful models for complex planning)
• Prompt: Copy the entire content from prompts/planner.md
• Context: Enable "Full folder context"
• Tools: Enable "Search codebase", "Read file", "Edit & Reapply" (to generate a feature PRD file)
• Automation: Disable auto-apply, auto-run, auto-fix (manual review required)

2. Executor mode

• Name: "Executor"
• Model: Claude Sonnet 4 or Gemini 2.5 Flash (fast, capable models)
• Prompt: Copy the entire content from prompts/executor.md
• Context: Enable "Full folder context"
• Tools: Enable all tools (file editing, terminal, etc.)
• Automation: Optional - Enable auto-apply edits, auto-run, and auto-fix errors for autonomous execution. Prefer manual review.

Kilo Code (modes)

1. Planner mode

• Name: "Planner"

• API Configuration: Gemini 2.5 Pro or Claude Opus 4 (powerful models for complex planning)

• Save Location: "Global" (available in all workspaces)

• Role Definition: Copy <persona> tag content from prompts/planner.md

• Short description (for humans): "Plan and design technical specification"

• When to Use (optional): "Use this mode when you need to plan, design, or strategize before implementation. Perfect for breaking down complex problems, creating technical specifications, designing system architecture, or brainstorming solutions before coding."

• Available Tools: Enable "Read Files", "Edit Files (Markdown files only)", "Use Browser"

  • Use this mode's group configuration to limit to markdown files only:

    groups:
      - read
      - - edit
        - fileRegex: \.md$
          description: Markdown files only
      - browser

• Custom Instructions (optional): Copy the entire content except <persona> tag from prompts/planner.md

2. Executor mode

• Name: "Executor"
• API Configuration: Claude Sonnet 4 or Gemini 2.5 Flash (fast, capable models)
• Save Location: "Global" (available in all workspaces)
• Role Definition: Copy <persona> tag content from prompts/executor.md
• Short description (for humans): "Implement features based on pre-approved specification"
• When to Use (optional): "Use this mode when you need to implement a feature whose spec you approved during the Planning mode. Ideal for implementing features, fixing bugs, creating new files, or making code improvements across any programming language or framework."
• Available Tools: Enable "Read Files", "Edit Files", "Use Browser", "Run Commands", "Use MCP"
• Custom Instructions (optional): Copy the entire content except <persona> tag from prompts/executor.md

GitHub Copilot (custom chat modes)

To edit modes, select Configure Modes... in the mode dropdown in the chat window.

Choose User Data Folder to make it available in all workspaces.

1. Planner mode

• Name: "Planner"

• Copy the content below into the markdown file that opens automatically:

  ---
  description: 'Plan and design technical specification'
  tools: ['codebase', 'usages', 'fetch', 'githubRepo', 'editFiles', 'search']
  ---
  
  Copy the entire content from `prompts/planner.md`.

2. Executor mode

• Name: "Executor"

• Copy the content below into the markdown file that opens automatically:

  ---
  description: 'Implement features based on pre-approved specification'
  tools: ['codebase', 'usages', 'vscodeAPI', 'problems', 'changes', 'testFailure', 'terminalSelection', 'terminalLastCommand', 'openSimpleBrowser', 'fetch', 'findTestFiles', 'searchResults', 'githubRepo', 'extensions', 'editFiles', 'runNotebooks', 'search', 'new', 'runCommands', 'runTasks']
  ---
  
  Copy the entire content from `prompts/executor.md`.

Check out how to set up custom modes in other AI assistant tools.

Workflow Usage

1. Switch to the Planner persona in your AI tool.
2. Provide a detailed feature description (core functionality requirements, expected user interactions, any technical constraints or dependencies, etc.).
3. The Planner will guide you through an interactive process to create a feature blueprint — prd.md in a new docs/specs/{feature-name}/ directory.
4. Review and approve to complete the plan.
5. Switch to the Executor persona.
6. The Executor will read relevant system documentation for context.
7. The Executor will automatically pick up the approved plan and implement it.

Other Approaches

• Kiro-style Prompts for Spec-Driven Development
• Two-step approach to AI coding
• 🚀 AI Dev Tasks 🤖
• Agent OS
• Prompts for plan & execute workflow in Cursor

Author and License

Nick Plekhanov, a full stack engineer.

CC0 1.0 Universal license.