Can I save the prompts generated by this skill?

Yes, by using the --save flag or responding to the prompt at the end of the workflow, you can save the generated instructions to .claude/prompts/ for future use.

What is the difference between this and the standard generate-prompt command?

While /generate-prompt only creates the instruction text, this skill adds an immediate execution step that implements the proposed changes as soon as you approve the prompt.

How does this skill analyze my codebase?

The skill scans for package managers, reads documentation files like README.md and CLAUDE.md, identifies relevant code via Grep, and checks recent Git history to build a complete context model.

Will it modify my code without permission?

No. The skill is designed with a safety-first approach. It always displays the generated prompt for your review and requires explicit confirmation before performing any file modifications.

Does it work with specific tech stacks?

It is tech-stack agnostic. It automatically detects your environment by looking for configuration files like package.json, Cargo.toml, pyproject.toml, and others.

Prompt Generator & Executor

Name: Prompt Generator & Executor
Author: jason-hchsieh

byjason-hchsieh

0•

Productivity & Workflow

Automates codebase analysis and task planning by generating high-fidelity prompts and executing them directly upon user approval.

This skill streamlines the complex development lifecycle by bridging the gap between planning and execution within Claude Code. It performs a deep dive into your repository—analyzing project structure, identifying relevant code patterns, and scanning Git history—to construct a comprehensive, context-aware instruction set. Unlike standard generation tools, it features an integrated execution loop that allows you to review the generated plan and immediately trigger the implementation, ensuring all changes are grounded in your project's specific architecture while maintaining full human oversight.

Key Features

01Context-rich prompt generation incorporating relevant file snippets and architectural patterns

02Automatic prompt persistence to save successful workflows for future reuse

030 GitHub stars

04Interactive confirmation loop to prevent unauthorized or unintended code execution

05Full-cycle task implementation based on generated requirements and hints

06Deep codebase analysis including tech stack detection and dependency mapping

Use Cases

01Expediting bug fixes by gathering diagnostic context and executing solutions in one flow

02Refactoring legacy code by automatically identifying all relevant dependencies and patterns

03Implementing complex features across multiple files while maintaining architectural consistency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jason-hchsieh/marketplace generate-and-execute

For use in Claude.ai and ChatGPT

Download Skill