agentsclimarketplace

New mcp prompt

Skill stevekinney/lesson-plan-adapter/.claude/skills/new-mcp-prompt

Install
npx -y skills add stevekinney/lesson-plan-adapter --skill new-mcp-prompt

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.8 KB, as published. Nobody here has run it

Skill: New MCP Prompt

Scaffold a new MCP prompt in the template.

Steps

  1. Create packages/mcp/src/prompts/<prompt-name>.ts (kebab-case filename)
  2. Define the prompt object with:
    • name: snake_case (e.g., my_prompt_name)
    • description: Clear, concise description of what the prompt does
    • arguments: Raw Zod shape (not wrapped in z.object()) defining the prompt arguments
    • handler: Async function with (arguments_: { ... }, context: { userId: string }) signature
  3. Follow the error handling pattern:
    • Create a child logger: logger.child({ prompt: 'prompt_name', userId: context.userId })
    • Wrap handler body in try/catch
    • On success: return { messages: [{ role: 'user', content: { type: 'text', text: '...' } }] }
    • On failure: log requestLogger.error({ err }, 'Prompt failed') and return a fallback message
    • Prompts must never throw — always return a structured MCP response
  4. Register in packages/mcp/src/server.ts:
    server.registerPrompt(
      myPrompt.name,
      { description: myPrompt.description, argsSchema: myPrompt.arguments },
      async (arguments_) => myPrompt.handler(arguments_, context),
    );
    
  5. Re-export from packages/mcp/src/index.ts

Template

import { z } from 'zod';
import { logger } from '../logger.js';

export const myPromptNamePrompt = {
  name: 'my_prompt_name' as const,
  description: 'Description of what this prompt does.',
  arguments: {
    myArgument: z.string().describe('Description of the argument'),
  },
  handler: async (arguments_: { myArgument: string }, context: { userId: string }) => {
    const requestLogger = logger.child({ prompt: 'my_prompt_name', userId: context.userId });
    try {
      requestLogger.info('Prompt requested');
      return {
        messages: [
          {
            role: 'user' as const,
            content: {
              type: 'text' as const,
              text: `Prompt text using ${arguments_.myArgument} for user ${context.userId}`,
            },
          },
        ],
      };
    } catch (error) {
      requestLogger.error({ err: error }, 'Prompt failed');
      return {
        messages: [
          {
            role: 'user' as const,
            content: {
              type: 'text' as const,
              text: 'An error occurred while generating the prompt.',
            },
          },
        ],
      };
    }
  },
};

Using Markdown Templates

Prompt handlers can import .md files as strings for template content:

import templateContent from './templates/my-template.md' with { type: 'text' };

// In the handler:
const rendered = templateContent.replace('{variable}', actualValue);

Create template files in packages/mcp/src/prompts/templates/. The text-imports.d.ts declaration in packages/mcp/src/ provides TypeScript support.

Keep looking

Skills are one crate of 608,475. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.