SSH MCP Server

/** * COMMAND STATE SYNCHRONIZATION EPIC - CRITICAL VALIDATION TEST * * This test validates that the epic requirements are fully met: * 1. Error code is BROWSER_COMMANDS_EXECUTED (not BROWSER_COMMANDS_ACTIVE) * 2. Command results are included for informed decision-making * 3. Claude Code can see both commands AND their execution results */ import { MCPSSHServer } from "../src/mcp-ssh-server.js"; import { SSHConnectionManager } from "../src/ssh-connection-manager.js"; describe('Command State Synchronization Epic - Critical Validation', () => { let mcpServer: MCPSSHServer; let sshManager: SSHConnectionManager; const testSessionName = 'epic-validation-session'; beforeEach(async () => { sshManager = new SSHConnectionManager(); mcpServer = new MCPSSHServer({}, sshManager); // Create real SSH connection using CLAUDE.md compliant approach (no mocks) const connectionResult = await mcpServer.callTool("ssh_connect", { name: testSessionName, host: "localhost", username: "jsbattig", keyFilePath: "/home/jsbattig/.ssh/id_ed25519" }); // Verify connection was successful if (!(connectionResult as any).success) { throw new Error(`Failed to establish SSH connection: ${(connectionResult as any).error}`); } }); afterEach(async () => { if (sshManager.hasSession(testSessionName)) { await mcpServer.callTool("ssh_disconnect", { sessionName: testSessionName }); } }); it('should return BROWSER_COMMANDS_EXECUTED error code (not BROWSER_COMMANDS_ACTIVE)', async () => { // Arrange: Add browser command to buffer sshManager.addBrowserCommand(testSessionName, 'ls -la', 'test-cmd-1', 'user'); // Act: Try to execute MCP command const result = await mcpServer.callTool("ssh_exec", { sessionName: testSessionName, command: "pwd" }); // Assert: Error code must be BROWSER_COMMANDS_EXECUTED expect(result).toMatchObject({ success: false, error: 'BROWSER_COMMANDS_EXECUTED', // CRITICAL: Must NOT be BROWSER_COMMANDS_ACTIVE message: 'User executed commands directly in browser', retryAllowed: true }); // Assert: Must be object array, not string array expect((result as any).browserCommands).toBeInstanceOf(Array); expect((result as any).browserCommands[0]).toBeInstanceOf(Object); expect((result as any).browserCommands[0]).toHaveProperty('command'); expect((result as any).browserCommands[0]).toHaveProperty('result'); }); it('should include command execution results for informed decision-making', async () => { // Arrange: Add browser command and simulate execution result sshManager.addBrowserCommand(testSessionName, 'pwd', 'executed-cmd-1', 'user'); // Simulate command execution with actual result sshManager.updateBrowserCommandResult(testSessionName, 'executed-cmd-1', { stdout: '/home/user/project', stderr: '', exitCode: 0 }); // Act: Try to execute MCP command const result = await mcpServer.callTool("ssh_exec", { sessionName: testSessionName, command: "ls" }); // Assert: Must include complete command execution results expect(result).toEqual({ success: false, error: 'BROWSER_COMMANDS_EXECUTED', message: 'User executed commands directly in browser', browserCommands: [ { command: 'pwd', commandId: 'executed-cmd-1', timestamp: expect.any(Number), source: 'user', result: { stdout: '/home/user/project', // CRITICAL: Must include actual command output stderr: '', exitCode: 0 // CRITICAL: Must include actual exit code } } ], retryAllowed: true }); }); it('should provide both successful and failed command results for complete context', async () => { // Arrange: Add multiple commands with different results sshManager.addBrowserCommand(testSessionName, 'pwd', 'success-cmd', 'user'); sshManager.addBrowserCommand(testSessionName, 'invalid-command', 'failed-cmd', 'user'); // Simulate successful command result sshManager.updateBrowserCommandResult(testSessionName, 'success-cmd', { stdout: '/home/user', stderr: '', exitCode: 0 }); // Simulate failed command result sshManager.updateBrowserCommandResult(testSessionName, 'failed-cmd', { stdout: '', stderr: 'command not found: invalid-command', exitCode: 127 }); // Act: Try to execute MCP command const result = await mcpServer.callTool("ssh_exec", { sessionName: testSessionName, command: "echo test" }); // Assert: Must include both success and failure results for informed decisions expect((result as any).browserCommands).toHaveLength(2); // Successful command result expect((result as any).browserCommands[0]).toEqual({ command: 'pwd', commandId: 'success-cmd', timestamp: expect.any(Number), source: 'user', result: { stdout: '/home/user', stderr: '', exitCode: 0 // Success } }); // Failed command result expect((result as any).browserCommands[1]).toEqual({ command: 'invalid-command', commandId: 'failed-cmd', timestamp: expect.any(Number), source: 'user', result: { stdout: '', stderr: 'command not found: invalid-command', exitCode: 127 // Failure } }); }); it('should show default result values for commands not yet executed', async () => { // Arrange: Add command to buffer but don't execute it sshManager.addBrowserCommand(testSessionName, 'pending-command', 'not-executed', 'user'); // Act: Try to execute MCP command immediately const result = await mcpServer.callTool("ssh_exec", { sessionName: testSessionName, command: "test" }); // Assert: Not-yet-executed commands show default result values expect((result as any).browserCommands[0]).toEqual({ command: 'pending-command', commandId: 'not-executed', timestamp: expect.any(Number), source: 'user', result: { stdout: '', stderr: '', exitCode: -1 // -1 indicates command not yet executed } }); }); });