A structured AI agent skill that converts vague project ideas into comprehensive technical specifications and step-by-step implementation plans through iterative refinement.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install pdd
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install pdd using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is pdd (Prompt-Driven Development)?
The pdd skill implements a rigorous Prompt-Driven Development methodology, serving as a bridge between initial concepts and actual code. By utilizing Openclaw Skills, this tool ensures that requirements are fully clarified, researched, and designed before implementation begins. It forces a disciplined, human-in-the-loop workflow where the agent and user collaborate to build a source of truth, reducing technical debt and preventing scope creep.
This skill is particularly effective for complex software tasks where architectural decisions and technology choices must be documented. By following the pdd process within the Openclaw Skills ecosystem, developers can generate high-quality artifacts that are ready for autonomous implementation by agents like Ralph or for manual development by a technical team.
pdd (Prompt-Driven Development) Use Cases
- Turning a high-level feature request into a detailed design document with Mermaid architecture diagrams.
- Conducting deep-dive research into specific libraries or APIs before starting a new module.
- Creating a TDD-ready implementation plan that breaks down a project into demoable, incremental steps.
- Standardizing the specification process for teams using AI coding assistants to ensure consistency across project artifacts.
How pdd (Prompt-Driven Development) Works
- The user provides a rough idea via text, file, or URL, and the skill initializes a project directory.
- The skill enters a Requirements Clarification phase, asking targeted questions one at a time to build a requirements.md file.
- A Research phase is initiated to explore technology choices, citing sources and documenting findings in specialized topic files.
- All gathered information is synthesized into a standalone design.md document covering architecture, data models, and error handling.
- An implementation plan is generated in plan.md, featuring a checklist of incremental steps with test requirements and integration notes.
- The process concludes with a summary and an optional integration step to hand off the project to autonomous implementation agents.
pdd (Prompt-Driven Development) Setup
To use the pdd skill, ensure your agent environment is configured to load Openclaw Skills. You can trigger the skill by providing the required parameters:
# Example parameters for the pdd skill
rough_idea: "Build a real-time notification system using WebSockets"
project_dir: "specs/notifications/" # Optional: defaults to kebab-case of idea
The agent will then guide you through the iterative steps, requiring explicit confirmation before moving between phases.
pdd (Prompt-Driven Development) Data Schema & Taxonomy
The skill organizes all output within a dedicated project directory using the following structure:
| File/Folder | Description |
|---|---|
rough-idea.md |
Stores the initial input or concept provided by the user. |
requirements.md |
A chronological record of the Q&A process used to define specifications. |
research/ |
A subdirectory containing individual markdown files for specific technical findings. |
design.md |
The primary technical specification including architecture and data models. |
plan.md |
A numbered implementation guide with a progress tracking checklist. |
summary.md |
A high-level overview of all generated artifacts and recommended next steps. |
PROMPT.md |
(Optional) A concise prompt for autonomous agents to begin implementation. |
pdd (Prompt-Driven Development) Advanced Features
- Iterative Phase Jumping: Users can move back and forth between requirements and research at any time to refine decisions.
- Mermaid Diagram Support: Automatically generates visual architecture and data flow diagrams within the design documents.
- Ralph Integration: Generates specialized configuration and prompt files to transition from design to autonomous code implementation.
- Given-When-Then Formatting: Defines acceptance criteria in a format suitable for machine verification and automated testing.
- Real-time Documentation: Findings and answers are appended to files immediately to ensure no data is lost during the session.
SKILL.md
Loading
Related Openclaw Skills
Find Code Tasks
A specialized utility for listing and filtering code tasks within a repository to maintain a clear overview of project status.
Evaluate Presets
A specialized testing framework for validating AI agent persona presets and auditing routing performance through automated CLI scripts.
Create Hat Collection
A conversational generator that transforms workflow ideas into validated, production-ready multi-agent YAML presets.
Code Task Generator
An intelligent utility that transforms raw project requirements or PDD implementation plans into structured, actionable Markdown task files for AI agents and developers.
playwriter
A browser automation tool that executes Playwright scripts against persistent, local Chrome sessions to preserve logins and extensions.
PR Demo Creation
A professional workflow for recording, converting, and validating terminal sessions into animated GIFs and SVGs for pull requests.
Featured*
