SystemPrompt Coding Agent

/** * @fileoverview MCP Server configuration and capabilities - Example Implementation * @module constants/server/server-config * * @remarks * This module defines the server metadata and capabilities for this example MCP server * implementation. It demonstrates how to properly configure a type-safe MCP server * that implements the full MCP specification including tools, prompts, resources, * sampling, and OAuth 2.1 authentication. * * This implementation provides a coding agent orchestrator that manages Claude Code CLI * and Gemini CLI sessions to perform coding tasks. * * @see {@link https://modelcontextprotocol.io/specification/2025-06-18/server | MCP Server Specification} */ import type { Implementation, ServerCapabilities } from "@modelcontextprotocol/sdk/types.js"; /** * Server implementation metadata * * @remarks * Provides identifying information about this example MCP server implementation. * This demonstrates proper type-safe configuration with full specification support. * * Key features demonstrated: * - Complete MCP specification implementation * - Type-safe TypeScript architecture * - Production-ready patterns and best practices * - Task orchestration for coding agents * - Integration with Claude Code and Gemini CLIs */ export const serverConfig: Implementation = { name: "systemprompt-coding-agent", version: "0.01", metadata: { name: "SystemPrompt Coding Agent", description: "MCP server for orchestrating Claude Code CLI and Gemini CLI sessions to perform coding tasks. " + "Provides tools for managing coding agents, task execution, and state persistence. " + "Perfect for automating complex coding workflows across multiple AI assistants.", icon: "code", color: "blue", serverStartTime: Date.now(), environment: process.env.NODE_ENV || "production", customData: { serverType: "coding-agent-orchestrator", implementationFeatures: [ "task-orchestration", "agent-management", "state-persistence", "elicitation", "structured-data", "notifications", "session-management" ], supportedAgents: ["claude-code-cli", "gemini-cli"], capabilities: "Orchestrates multiple AI coding assistants to complete complex tasks", }, }, }; /** * Server capabilities declaration * * @remarks * This declares all MCP capabilities that this example server supports. * Each capability corresponds to a feature in the MCP specification: * * - tools: Interactive functions the AI can call * - prompts: Predefined prompt templates * - resources: Dynamic content the AI can read * - logging: Server-side logging capability * * This example implements ALL capabilities to serve as a complete reference. * */ export const serverCapabilities: { capabilities: ServerCapabilities } = { capabilities: { /** Full tool support with type-safe handlers */ tools: {}, /** Dynamic prompt generation */ prompts: {}, /** Resource listing and reading */ resources: {}, /** Client-requested logging support */ logging: {}, /** Filesystem roots support */ roots: { /** We support roots/list_changed notifications */ listChanged: true }, /** Resource templates support for dynamic URIs */ resourceTemplates: {}, }, }; /** * Additional server configuration constants */ export const SERVER_CONFIG = { /** Session timeout in milliseconds (30 minutes) */ SESSION_TIMEOUT: 30 * 60 * 1000, /** Maximum concurrent sessions */ MAX_SESSIONS: 100, /** Rate limiting configuration */ RATE_LIMIT: { /** Rate limit window in milliseconds (1 minute) */ windowMs: 60000, /** Maximum requests per window */ maxRequests: 100, }, /** Maximum request body size (10MB) */ MAX_REQUEST_SIZE: 10 * 1024 * 1024, /** MCP protocol version */ PROTOCOL_VERSION: "2025-06-18", /** MCP SDK version */ SDK_VERSION: "@modelcontextprotocol/sdk@1.13.0", } as const;