Claude Code General-Purpose Prompt Collection

English prompt templates you can copy and use immediately when working with external resources (MCP / repos / Skills / Hooks / Agents) in Claude Code.

MCP

Install New MCP Server

MCP server description → prerequisites → settings.json config → test commands → usage examples

Repo

Analyze External GitHub Repo

Comprehensive prompt for analyzing repo purpose, stack, structure, usage steps, variation scenarios, and license

Skill

Write Custom Skill

Purpose-based SKILL.md writing → trigger conditions → step-by-step instructions → execution examples

Hooks

Configure Hooks

Select event → set matcher → generate settings.json → verify behavior

Agents

Configure Sub-agent Pattern

Role separation → task scope → communication method → prompt chain → error recovery

Integrated

Repo + MCP + Hooks Integrated Setup

After clone: write CLAUDE.md → recommend MCP → configure Hooks → add skills → workflow

---

1. Install New MCP Server

I want to connect the [MCP server name] MCP server to Claude Code.
1. Explain in one line what this MCP server does
2. Tell me the prerequisites for installation
3. Show the exact JSON config to add to .claude/settings.json
4. Give me a test command to verify it works after installation
5. Give me examples of how to use it in a university project

2. Analyze External GitHub Repo

Please analyze this GitHub repository: [repo URL]
1. Summarize the purpose and core features in one line
2. List the tech stack (language, framework, dependencies)
3. Explain the directory structure and key files
4. Provide a step-by-step guide for using it after cloning in Claude Code
5. Give 3 scenarios where a university student could modify this repo for their own use
6. Check the license and any points to be aware of

3. Write Custom Skill

I want to create my own Claude Code Skill.
Purpose: [specific purpose]
1. Write a SKILL.md file in .claude/skills/
2. Clearly describe trigger conditions in the description
3. Include step-by-step instructions in the body
4. Provide 3 example execution prompts
5. Explain how to test it

4. Configure Hooks

I want to configure Claude Code hooks.
Purpose: [specific purpose]
1. Select the appropriate hook event (PreToolUse/PostToolUse/Notification, etc.)
2. Set the matcher pattern
3. Generate the exact JSON to add to .claude/settings.json
4. How to test that it's working
5. Possible issues when misconfigured and how to resolve them

5. Configure Sub-agent Pattern

I want to use the sub-agent pattern in Claude Code.
Project context: [project description]
1. Separate main/sub agent roles
2. Define each agent's task scope
3. Agent-to-agent communication / result passing method
4. Prompt chain for actual execution
5. Recovery strategy when errors occur

6. Repo + MCP + Hooks Integrated Setup

I want to clone [GitHub repo URL] and develop with Claude Code.
Please set up an optimized Claude Code environment for this project:
1. Write a CLAUDE.md file tailored to the project
2. Recommend necessary MCP servers + settings.json configuration
3. Configure code quality hooks (formatting, linting, testing)
4. Suggest useful custom Skills to add
5. Step-by-step development workflow summary

---

Field	Value
License	MIT
Translation Date	2026-04-12
Author	Claude-Code-Study Project

Replace the [bracketed] parts of each prompt with your own situation before using.