Node Code Sandbox MCP

Instructions

Implementation Reference

• src/tools/aiGenerate.ts:29-80 (handler)

  The async handler function that implements the ai_generate tool by calling the Google Gemini API with the provided prompt, model, and maxTokens.

  export default async function aiGenerate({
    prompt,
    model = 'models/gemini-2.0-flash-exp',
    maxTokens,
  }: {
    prompt: string;
    model?: string;
    maxTokens?: number;
  }): Promise<McpResponse> {
    const apiKey = process.env.GEMINI_API_KEY;
    if (!apiKey) {
      logger.error('GEMINI_API_KEY is not set in environment variables');
      return { content: [textContent('Error: Gemini API key not configured.')] };
    }
    try {
      const response = await fetch(
        'https://generativelanguage.googleapis.com/v1beta/models/' +
          encodeURIComponent(model) +
          ':generateContent?key=' +
          apiKey,
        {
          method: 'POST',
          headers: { 'Content-Type': 'application/json' },
          body: JSON.stringify({
            contents: [{ parts: [{ text: prompt }] }],
            ...(maxTokens
              ? { generationConfig: { maxOutputTokens: maxTokens } }
              : {}),
          }),
        }
      );
      if (!response.ok) {
        const errorText = await response.text();
        logger.error('Gemini API error', errorText);
        return { content: [textContent('Gemini API error: ' + errorText)] };
      }
      const data = await response.json();
      const text =
        data.candidates?.[0]?.content?.parts?.[0]?.text || '[No response]';
      return { content: [textContent(text)] };
    } catch (error) {
      logger.error('Failed to call Gemini API', error);
      return {
        content: [
          textContent(
            'Error calling Gemini: ' +
              (error instanceof Error ? error.message : String(error))
          ),
        ],
      };
    }
  }

• src/tools/aiGenerate.ts:11-19 (schema)

  Zod schema defining the input arguments for the ai_generate tool: prompt (required), model (optional), maxTokens (optional).

  export const argSchema = {
    prompt: z.string().min(1).describe('Prompt to send to Gemini'),
    model: z
      .string()
      .optional()
      .default('models/gemini-2.0-flash-exp')
      .describe('Gemini model name'),
    maxTokens: z.number().optional().describe('Maximum tokens in the response'),
  };

• src/server.ts:115-120 (registration)

  Registration of the 'ai_generate' tool on the MCP server, specifying name, description, argSchema, and handler from aiGenerate.ts.

  server.tool(
    'ai_generate',
    'Generate text using Google Gemini. Provide a prompt and optional model name.',
    (await import('./tools/aiGenerate.ts')).argSchema,
    (await import('./tools/aiGenerate.ts')).default
  );