Add MCP Prompt

Add a new prompt template to an existing FastMCP server for reusable LLM interactions.

Usage

/add-mcp-prompt [prompt-name]

Arguments

• [prompt-name]: Optional - Name for the prompt (will prompt if not provided)

Execution Instructions for Claude Code

When this command is run:

1. Locate the FastMCP server in the current project
2. Gather prompt requirements through interactive questions
3. Generate prompt code based on configuration
4. Add prompt to server by modifying existing files

Interactive Session Flow

1. Locate Server

Looking for FastMCP server...

Found: src/server.ts

Is this the correct server file? (yes/no):

2. Prompt Basics

Let's add a new prompt to your MCP server.

Prompt name (lowercase, use hyphens):
Example: "git-commit", "code-review", "summarize-document"

Prompt name:

Prompt description (shown to clients):
Example: "Generate a Git commit message from changes"

Description:

3. Prompt Arguments

Does this prompt require arguments?

1. No arguments
2. Single required argument
3. Multiple arguments (some optional)

Select (1-3):

If arguments needed:

Let's define the arguments.

Argument 1:
  Name:
  Description:
  Required? (yes/no):

Add another argument? (yes/no):

4. Argument Completion

For each argument:

How should "{argName}" support auto-completion?

1. No completion
2. Enum - Fixed list of options
3. Dynamic completion - Custom logic

Select (1-3):

If enum:

Enter enum values (comma-separated):
Example: "typescript,javascript,python,go"

Values:

If dynamic:

What kind of dynamic completion?

1. Search/filter from list
2. API lookup
3. File system suggestions
4. Custom (I'll implement)

Select (1-4):

5. Prompt Template

How would you like to define the prompt template?

1. Simple text template with argument interpolation
2. Multi-part structured prompt
3. I'll write the template manually

Select (1-3):

If simple template:

Write your prompt template. Use {argumentName} for argument placeholders.

Example:
"Review the following code and provide feedback on:
- Code quality
- Potential bugs
- Performance issues

Code to review:
{code}"

Your template:

6. Prompt Options

Additional options:

Should the prompt return multiple messages? (yes/no) [default: no]:

Include system message? (yes/no) [default: no]:

Generated Code Examples

Simple Prompt (No Arguments)

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  load: async () => {
    return "Your prompt text here";
  },
});

Prompt with Single Argument

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  arguments: [
    {
      name: "{{argName}}",
      description: "{{argDescription}}",
      required: true,
    },
  ],
  load: async (args) => {
    return "{{promptTemplate}}".replace("{{{argName}}}", args.{{argName}});
  },
});

Prompt with Multiple Arguments

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  arguments: [
    {{#each arguments}}
    {
      name: "{{name}}",
      description: "{{description}}",
      required: {{required}},
    },
    {{/each}}
  ],
  load: async (args) => {
    const { {{argumentNames}} } = args;

    return "{{promptTemplate}}";
  },
});

Prompt with Enum Completion

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  arguments: [
    {
      name: "{{argName}}",
      description: "{{argDescription}}",
      required: true,
      enum: [{{#each enumValues}}"{{this}}"{{#unless @last}}, {{/unless}}{{/each}}],
    },
  ],
  load: async (args) => {
    return "Process the following in " + args.{{argName}} + " format:\n\n{content}";
  },
});

Prompt with Dynamic Completion

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  arguments: [
    {
      name: "{{argName}}",
      description: "{{argDescription}}",
      required: true,
      complete: async (value) => {
        // Search/filter logic
        const options = await getAvailableOptions();
        const filtered = options.filter(opt =>
          opt.toLowerCase().includes(value.toLowerCase())
        );
        return { values: filtered };
      },
    },
  ],
  load: async (args) => {
    const data = await loadData(args.{{argName}});
    return "Analyze the following data:\n\n" + JSON.stringify(data, null, 2);
  },
});

Git Commit Message Prompt

server.addPrompt({
  name: "git-commit",
  description: "Generate a Git commit message from changes",
  arguments: [
    {
      name: "changes",
      description: "Git diff or description of changes",
      required: true,
    },
    {
      name: "type",
      description: "Commit type",
      required: false,
      enum: ["feat", "fix", "docs", "style", "refactor", "test", "chore"],
    },
  ],
  load: async (args) => {
    const typeHint = args.type
      ? "Use commit type: " + args.type
      : "Choose appropriate commit type (feat, fix, docs, etc.)";

    return "Generate a concise Git commit message following conventional commits format.\n\n" +
      typeHint + "\n\n" +
      "Changes:\n" + args.changes + "\n\n" +
      "Format: <type>(<optional scope>): <description>\n\n" +
      "Keep the description under 72 characters.";
  },
});

Code Review Prompt

server.addPrompt({
  name: "code-review",
  description: "Review code for quality, bugs, and improvements",
  arguments: [
    {
      name: "code",
      description: "Code to review",
      required: true,
    },
    {
      name: "language",
      description: "Programming language",
      required: false,
      enum: ["typescript", "javascript", "python", "go", "rust"],
    },
    {
      name: "focus",
      description: "Focus area for review",
      required: false,
      enum: ["security", "performance", "readability", "all"],
    },
  ],
  load: async (args) => {
    const lang = args.language || "the detected language";
    const focus = args.focus || "all aspects";

    return "Review the following " + lang + " code, focusing on " + focus + ".\n\n" +
      "Provide feedback on:\n" +
      "1. Code quality and best practices\n" +
      "2. Potential bugs or edge cases\n" +
      "3. Performance considerations\n" +
      "4. Security issues (if applicable)\n" +
      "5. Suggestions for improvement\n\n" +
      "Code:\n```\n" + args.code + "\n```\n\n" +
      "Format your response with clear sections and specific line references where applicable.";
  },
});

Document Summary Prompt

server.addPrompt({
  name: "summarize",
  description: "Summarize a document or text",
  arguments: [
    {
      name: "content",
      description: "Content to summarize",
      required: true,
    },
    {
      name: "length",
      description: "Summary length",
      required: false,
      enum: ["brief", "detailed", "bullet-points"],
    },
    {
      name: "audience",
      description: "Target audience",
      required: false,
    },
  ],
  load: async (args) => {
    const style = {
      brief: "Provide a 2-3 sentence summary",
      detailed: "Provide a comprehensive summary with key points",
      "bullet-points": "Summarize as a bulleted list of key points",
    }[args.length || "brief"];

    const audienceNote = args.audience
      ? "Write for: " + args.audience
      : "";

    return style + ". " + audienceNote + "\n\n" +
      "Content to summarize:\n" + args.content;
  },
});

Multi-Message Prompt

server.addPrompt({
  name: "{{name}}",
  description: "{{description}}",
  arguments: [
    {
      name: "context",
      description: "Context for the conversation",
      required: true,
    },
  ],
  load: async (args) => {
    // Return array of messages for multi-turn context
    return [
      {
        role: "user",
        content: {
          type: "text",
          text: "Context: " + args.context,
        },
      },
      {
        role: "assistant",
        content: {
          type: "text",
          text: "I understand the context. How can I help?",
        },
      },
      {
        role: "user",
        content: {
          type: "text",
          text: "Now proceed with the main task...",
        },
      },
    ];
  },
});

File Placement

Where should the prompt code be added?

1. Inline in server.ts
2. New file in prompts/ directory
3. Existing prompts file

Select (1-3):

Implementation Notes

1. Find existing server: Locate FastMCP server files
2. Check for prompts directory: Create if needed
3. Generate code: Create prompt with selected options
4. Add imports: Ensure necessary imports exist
5. Insert code: Add prompt at appropriate location
6. Summary: Show generated prompt and usage example