read_multiple_files

Process multiple files simultaneously using glob patterns in local code repositories. Extract and analyze file contents for AI-driven project insights without manual uploads.

Instructions

批量读取多个文件内容，支持glob模式

Input Schema

TableJSON Schema

Name	Required	Description	Default
`maxFiles`	No	最大文件数量限制
`patterns`	Yes	文件模式数组，如 ['modules//services/.ts', 'config/*.ts']

Implementation Reference

src/index.ts:398-434 (handler)

Handler function that implements the logic for reading multiple files using glob patterns across synced directories, concatenating their contents with headers.

async ({ patterns, maxFiles = 20 }) => {
  let allContent = "批量文件内容:\n" + "=".repeat(50) + "\n";
  let fileCount = 0;

  for (const [prefix, absolutePath] of pathRegistry.entries()) {
    for (const pattern of patterns) {
      try {
        const fullPattern = path.join(absolutePath, pattern);
        const matchedFiles = await glob(fullPattern, {
          ignore: ['**/node_modules/**', '**/.git/**']
        });

        for (const file of matchedFiles) {
          if (fileCount >= maxFiles) break;

          try {
            const relativePath = path.relative(absolutePath, file);
            const content = await fs.readFile(file, 'utf-8');
            allContent += `\n📄 ${prefix}/${relativePath}\n`;
            allContent += "-".repeat(30) + "\n";
            allContent += content + "\n";
            fileCount++;
          } catch (error) {
            allContent += `\n无法读取: ${file}\n`;
          }
        }
      } catch (error) {
        console.error(`处理模式 ${pattern} 时出错:`, error);
      }

      if (fileCount >= maxFiles) break;
    }
    if (fileCount >= maxFiles) break;
  }

  return { content: [{ type: "text", text: allContent }] };
}

src/index.ts:394-397 (schema)

Zod schema defining the input parameters: patterns (array of glob patterns) and optional maxFiles.

{
  patterns: z.array(z.string()).describe("文件模式数组，如 ['modules/*/services/*.ts', 'config/*.ts']"),
  maxFiles: z.number().optional().default(20).describe("最大文件数量限制"),
},

src/index.ts:391-435 (registration)

Registration of the 'read_multiple_files' tool on the MCP server, including name, description, schema, and handler function.

server.tool(
  "read_multiple_files",
  "批量读取多个文件内容，支持glob模式",
  {
    patterns: z.array(z.string()).describe("文件模式数组，如 ['modules/*/services/*.ts', 'config/*.ts']"),
    maxFiles: z.number().optional().default(20).describe("最大文件数量限制"),
  },
  async ({ patterns, maxFiles = 20 }) => {
    let allContent = "批量文件内容:\n" + "=".repeat(50) + "\n";
    let fileCount = 0;

    for (const [prefix, absolutePath] of pathRegistry.entries()) {
      for (const pattern of patterns) {
        try {
          const fullPattern = path.join(absolutePath, pattern);
          const matchedFiles = await glob(fullPattern, {
            ignore: ['**/node_modules/**', '**/.git/**']
          });

          for (const file of matchedFiles) {
            if (fileCount >= maxFiles) break;

            try {
              const relativePath = path.relative(absolutePath, file);
              const content = await fs.readFile(file, 'utf-8');
              allContent += `\n📄 ${prefix}/${relativePath}\n`;
              allContent += "-".repeat(30) + "\n";
              allContent += content + "\n";
              fileCount++;
            } catch (error) {
              allContent += `\n无法读取: ${file}\n`;
            }
          }
        } catch (error) {
          console.error(`处理模式 ${pattern} 时出错:`, error);
        }

        if (fileCount >= maxFiles) break;
      }
      if (fileCount >= maxFiles) break;
    }

    return { content: [{ type: "text", text: allContent }] };
  }
);

Local Project Sync