mcp-adr-analysis-server

Instructions

Implementation Reference

• src/tools/get-server-context-tool.ts:42-87 (handler)

  The main handler function that executes the get_server_context tool. It creates a ServerContextGenerator instance, generates the current server context using provided managers, optionally writes it to .mcp-server-context.md, and returns the appropriate MCP CallToolResult.

  export async function getServerContext(
    args: GetServerContextArgs,
    kgManager: KnowledgeGraphManager,
    memoryManager: MemoryEntityManager,
    conversationManager: ConversationMemoryManager
  ): Promise<CallToolResult> {
    const { writeToFile = true, outputPath, includeDetailed = true, maxRecentItems = 5 } = args;
  
    const generator = new ServerContextGenerator();
  
    try {
      // Generate context
      const contextContent = await generator.generateContext(
        kgManager,
        memoryManager,
        conversationManager,
        { includeDetailed, maxRecentItems }
      );
  
      // Optionally write to file
      if (writeToFile) {
        await generator.writeContextFile(kgManager, memoryManager, conversationManager, outputPath);
      }
  
      return {
        content: [
          {
            type: 'text',
            text: writeToFile
              ? `✅ Server context updated and written to \`.mcp-server-context.md\`\n\nYou can now \`@.mcp-server-context.md\` to reference this context in conversations.\n\n---\n\n${contextContent}`
              : contextContent,
          },
        ],
      };
    } catch (error) {
      return {
        content: [
          {
            type: 'text',
            text: `❌ Failed to generate server context: ${error instanceof Error ? error.message : String(error)}`,
          },
        ],
        isError: true,
      };
    }
  }

• src/tools/get-server-context-tool.ts:92-120 (schema)

  MCP tool metadata including name, description, and inputSchema for registration and validation.

  export const getServerContextMetadata = {
    name: 'get_server_context',
    description:
      "Generate a comprehensive context file showing the server's current state, memory, and capabilities. Creates .mcp-server-context.md that can be @ referenced in conversations to give LLMs instant awareness of the server.",
    inputSchema: {
      type: 'object',
      properties: {
        writeToFile: {
          type: 'boolean',
          description: 'Whether to write the context to .mcp-server-context.md file',
          default: true,
        },
        outputPath: {
          type: 'string',
          description: 'Custom output path for the context file',
        },
        includeDetailed: {
          type: 'boolean',
          description: 'Include detailed information',
          default: true,
        },
        maxRecentItems: {
          type: 'number',
          description: 'Maximum number of recent items to show',
          default: 5,
        },
      },
    },
  };

• src/tools/get-server-context-tool.ts:14-37 (schema)

  TypeScript interface defining the input arguments for the handler function, matching the inputSchema.

  export interface GetServerContextArgs {
    /**
     * Whether to write the context to .mcp-server-context.md file
     * @default true
     */
    writeToFile?: boolean;
  
    /**
     * Custom output path for the context file
     */
    outputPath?: string;
  
    /**
     * Include detailed information
     * @default true
     */
    includeDetailed?: boolean;
  
    /**
     * Maximum number of recent items to show
     * @default 5
     */
    maxRecentItems?: number;
  }

• src/tools/tool-catalog.ts:1322-1339 (registration)

  Registration of the tool in the central TOOL_CATALOG used for dynamic discovery, ListTools responses via tool-dispatcher, and categorization.

  TOOL_CATALOG.set('get_server_context', {
    name: 'get_server_context',
    shortDescription: 'Get server context',
    fullDescription: 'Gets the current server context and configuration.',
    category: 'utility',
    complexity: 'simple',
    tokenCost: { min: 200, max: 500 },
    hasCEMCPDirective: true, // Phase 4.3: Simple tool - context retrieval
    relatedTools: ['check_ai_execution_status', 'manage_cache'],
    keywords: ['server', 'context', 'config', 'status'],
    requiresAI: false,
    inputSchema: {
      type: 'object',
      properties: {
        includeConfig: { type: 'boolean', default: false },
      },
    },
  });

• src/utils/server-context-generator.ts:90-109 (helper)

  Key helper method writeContextFile used by the handler to persist the generated context to disk and ensure gitignore entry.

  async writeContextFile(
    kgManager: KnowledgeGraphManager,
    memoryManager: MemoryEntityManager,
    conversationManager: ConversationMemoryManager,
    filePath?: string
  ): Promise<void> {
    const outputPath = filePath || path.join(this.config.projectPath, '.mcp-server-context.md');
    const content = await this.generateContext(kgManager, memoryManager, conversationManager);
  
    await fs.writeFile(outputPath, content, 'utf-8');
  
    // Ensure .mcp-server-context.md is in .gitignore (auto-generated files shouldn't be tracked)
    await this.ensureGitignoreEntry(this.config.projectPath, '.mcp-server-context.md');
  
    this.logger.info('Server context file updated', 'ServerContextGenerator', {
      path: outputPath,
      size: content.length,
    });
  }