mycontext-cli 2.0.8 → 2.0.9

package/dist/agents/implementations/ClaudeAgentWorkflow.js

@@ -1,661 +0,0 @@
-"use strict";
-/**
- * ClaudeAgentWorkflow Implementation
- *
- * Enhanced workflow agent that leverages Claude Agent SDK for advanced
- * context management, tool permissions, and MCP integration.
- */
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.ClaudeAgentWorkflow = void 0;
-const claudeAgentClient_1 = require("../../utils/claudeAgentClient");
-const contextManager_1 = require("../../utils/contextManager");
-const toolPermissions_1 = require("../../utils/toolPermissions");
-const mcpAgentIntegration_1 = require("../../utils/mcpAgentIntegration");
-const chalk_1 = __importDefault(require("chalk"));
-class ClaudeAgentWorkflow {
-    constructor(workingDirectory) {
-        this.name = "ClaudeAgentWorkflow";
-        this.description = "Enhanced workflow agent using Claude Agent SDK with advanced context management and MCP integration";
-        this.personality = "Advanced AI coordinator with deep project understanding and intelligent workflow orchestration";
-        this.llmProvider = "claude-agent-sdk";
-        this.expertise = [
-            "workflow-management",
-            "context-management",
-            "tool-permissions",
-            "mcp-integration",
-            "agent-coordination",
-            "error-handling",
-            "user-experience",
-        ];
-        this.completedSteps = new Set();
-        this.failedSteps = [];
-        this.userInteractions = 0;
-        this.totalRetries = 0;
-        this.workflowContext = {};
-        this.claudeAgentClient = new claudeAgentClient_1.ClaudeAgentClient(workingDirectory);
-        this.contextManager = new contextManager_1.ContextManager(workingDirectory);
-        this.permissionManager = new toolPermissions_1.ToolPermissionManager(workingDirectory);
-        this.mcpIntegration = new mcpAgentIntegration_1.MCPAgentIntegration(workingDirectory);
-    }
-    /**
-     * Check if Claude Agent SDK is available
-     */
-    isClaudeAgentAvailable() {
-        return this.claudeAgentClient.hasApiKey();
-    }
-    async run(input) {
-        const startTime = Date.now();
-        const maxRetries = input.maxRetries || 3;
-        console.log(chalk_1.default.blue.bold("🚀 Starting Claude Agent SDK Enhanced Workflow\n"));
-        console.log(chalk_1.default.gray(`Project: ${input.projectName}`));
-        console.log(chalk_1.default.gray(`Framework: ${input.framework}`));
-        console.log(chalk_1.default.gray(`Interactive Mode: ${input.interactive ? "Enabled" : "Disabled"}\n`));
-        try {
-            // Initialize Claude Agent SDK
-            await this.initializeClaudeAgent();
-            // Load and prepare context
-            await this.prepareWorkflowContext(input);
-            // Define enhanced workflow steps
-            const steps = this.defineEnhancedWorkflowSteps(input);
-            // Execute workflow with Claude Agent SDK
-            await this.executeEnhancedWorkflow(steps, maxRetries, input.interactive || false);
-            const duration = Date.now() - startTime;
-            const completedSteps = Array.from(this.completedSteps);
-            console.log(chalk_1.default.green.bold("\n✅ Claude Agent SDK Workflow completed successfully!"));
-            console.log(chalk_1.default.blue(`⏱️  Total duration: ${Math.round(duration / 1000)}s`));
-            console.log(chalk_1.default.blue(`📋 Completed steps: ${completedSteps.length}`));
-            console.log(chalk_1.default.blue(`🔄 Total retries: ${this.totalRetries}`));
-            console.log(chalk_1.default.blue(`💬 User interactions: ${this.userInteractions}`));
-            // Save final context
-            await this.contextManager.saveContext(this.workflowContext);
-            return {
-                success: true,
-                projectPath: input.projectName,
-                completedSteps,
-                failedSteps: this.failedSteps,
-                userInteractions: this.userInteractions,
-                totalRetries: this.totalRetries,
-                duration,
-            };
-        }
-        catch (error) {
-            const duration = Date.now() - startTime;
-            console.log(chalk_1.default.red.bold("\n❌ Claude Agent SDK Workflow failed"));
-            console.log(chalk_1.default.red(`Error: ${error instanceof Error ? error.message : "Unknown error"}`));
-            return {
-                success: false,
-                projectPath: input.projectName,
-                completedSteps: Array.from(this.completedSteps),
-                failedSteps: this.failedSteps,
-                userInteractions: this.userInteractions,
-                totalRetries: this.totalRetries,
-                duration,
-            };
-        }
-        finally {
-            // Cleanup
-            await this.claudeAgentClient.cleanup();
-        }
-    }
-    /**
-     * Initialize Claude Agent SDK with enhanced configuration
-     */
-    async initializeClaudeAgent() {
-        console.log(chalk_1.default.yellow("🔧 Initializing Claude Agent SDK with advanced features..."));
-        try {
-            // Check if Claude Agent SDK is available
-            if (!this.claudeAgentClient.hasApiKey()) {
-                throw new Error("Claude API key not configured. Set MYCONTEXT_CLAUDE_API_KEY or ANTHROPIC_API_KEY");
-            }
-            // Initialize MCP integration
-            await this.mcpIntegration.initialize(this.claudeAgentClient);
-            // Get permission configuration
-            const permissionConfig = this.permissionManager.getClaudeAgentConfig();
-            // Initialize Claude Agent with Phase 2 enhanced configuration
-            await this.claudeAgentClient.initialize({
-                model: "claude-3-5-sonnet-20241022",
-                temperature: 0.2,
-                maxTokens: 8000,
-                systemPrompt: this.getEnhancedSystemPrompt(),
-                ...permissionConfig,
-                contextManagement: true,
-                // NEW: Setting sources for reproducible builds
-                settingSources: ['project', 'local'],
-                // NEW: Register workflow-specific hooks
-                hooks: {
-                    PreToolUse: async (input, toolUseID, options) => {
-                        const toolInput = input;
-                        console.log(chalk_1.default.gray(`   🔧 Using tool: ${toolInput.tool_name || 'unknown'}`));
-                        return { continue: true };
-                    },
-                    PostToolUse: async (input, toolUseID, options) => {
-                        const toolInput = input;
-                        console.log(chalk_1.default.gray(`   ✅ Tool completed: ${toolInput.tool_name || 'unknown'}`));
-                        return { continue: true };
-                    },
-                    SessionStart: async (input, toolUseID, options) => {
-                        console.log(chalk_1.default.blue('   🚀 Agent session started'));
-                        return { continue: true };
-                    },
-                },
-                // NEW: Permission callback for interactive approval
-                canUseTool: async (toolName, input, options) => {
-                    // Auto-approve safe read operations
-                    if (['Read', 'Glob', 'Grep'].includes(toolName)) {
-                        return { behavior: 'allow', updatedInput: input };
-                    }
-                    // Always approve our custom MCP tools
-                    if ([
-                        'AnalyzeComponent',
-                        'ValidatePRD',
-                        'CheckTypes',
-                        'GenerateDocs',
-                        'DetectExistingComponents',
-                        'MapComponentsToRoutes',
-                        'GenerateScaffolding',
-                    ].includes(toolName)) {
-                        return { behavior: 'allow', updatedInput: input };
-                    }
-                    // For write operations, ask for permission (simplified for now)
-                    return { behavior: 'allow', updatedInput: input };
-                },
-            });
-            console.log(chalk_1.default.green("✅ Claude Agent SDK initialized with advanced features"));
-            // Log available agents
-            const availableAgents = this.claudeAgentClient.getAvailableAgents();
-            console.log(chalk_1.default.gray(`   📦 Available agents: ${availableAgents.length}`));
-        }
-        catch (error) {
-            console.log(chalk_1.default.red(`❌ Claude Agent SDK initialization failed: ${error.message}`));
-            throw error;
-        }
-    }
-    /**
-     * Prepare workflow context
-     */
-    async prepareWorkflowContext(input) {
-        console.log(chalk_1.default.yellow("📋 Preparing workflow context..."));
-        try {
-            // Load existing context
-            this.workflowContext = await this.contextManager.loadContext();
-            // Add workflow-specific context
-            this.workflowContext = await this.contextManager.mergeContext(this.workflowContext, {
-                userPrompt: input.description,
-                workingDirectory: input.projectName === "." ? process.cwd() : input.projectName,
-                previousOutputs: {
-                    workflowStart: {
-                        projectName: input.projectName,
-                        framework: input.framework,
-                        description: input.description,
-                        interactive: input.interactive,
-                        timestamp: new Date().toISOString(),
-                    },
-                },
-            });
-            // Check if context needs compaction
-            const contextStats = this.contextManager.getContextStats(this.workflowContext);
-            if (contextStats.totalSize > 50000) {
-                // 50KB limit
-                console.log(chalk_1.default.yellow("⚠️ Context size exceeds limit, compacting..."));
-                this.workflowContext = await this.contextManager.compactContext(this.workflowContext);
-            }
-            console.log(chalk_1.default.green("✅ Workflow context prepared"));
-        }
-        catch (error) {
-            console.log(chalk_1.default.yellow(`⚠️ Context preparation warning: ${error.message}`));
-        }
-    }
-    /**
-     * Define enhanced workflow steps with Claude Agent SDK
-     */
-    defineEnhancedWorkflowSteps(input) {
-        return [
-            {
-                id: "context-analysis",
-                name: "Context Analysis",
-                description: "Analyze project context and requirements using Claude Agent SDK",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "context-analysis",
-                    projectContext: this.workflowContext,
-                    requirements: input.description,
-                },
-                required: true,
-                retryable: true,
-                interactive: false,
-            },
-            {
-                id: "project-setup",
-                name: "Project Setup",
-                description: "Initialize project structure with enhanced context awareness",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "project-setup",
-                    projectName: input.projectName,
-                    framework: input.framework,
-                    context: this.workflowContext,
-                },
-                dependencies: ["context-analysis"],
-                required: true,
-                retryable: true,
-                interactive: false,
-            },
-            // NEW: Component Detection Step (Component-First Workflow)
-            {
-                id: "component-detection",
-                name: "Component Detection",
-                description: "Detect existing validated components for component-first workflow",
-                agentName: "component-first-builder",
-                agent: this,
-                input: {
-                    workflowType: "component-detection",
-                    projectPath: input.projectName,
-                    includeValidation: true,
-                },
-                dependencies: ["project-setup"],
-                required: true,
-                retryable: true,
-                interactive: false,
-            },
-            {
-                id: "requirement-gathering",
-                name: "Requirement Gathering",
-                description: "Gather detailed requirements using Claude Agent SDK",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "requirement-gathering",
-                    context: this.workflowContext,
-                    interactive: input.interactive,
-                },
-                dependencies: ["component-detection"],
-                required: input.interactive,
-                retryable: true,
-                interactive: true,
-            },
-            {
-                id: "context-generation",
-                name: "Context Generation",
-                description: "Generate comprehensive context files using Claude Agent SDK",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "context-generation",
-                    context: this.workflowContext,
-                    requirements: input.description,
-                },
-                dependencies: ["requirement-gathering"],
-                required: true,
-                retryable: true,
-                interactive: false,
-                // Skip if components already exist and should be reused
-                skip: (context) => context.componentDetection?.recommendation === "REUSE_COMPONENTS",
-            },
-            // NEW: Component Mapping Step (Only for REUSE_COMPONENTS mode)
-            {
-                id: "component-mapping",
-                name: "Component Mapping",
-                description: "Map existing components to routes intelligently",
-                agentName: "component-first-builder",
-                agent: this,
-                input: {
-                    workflowType: "component-mapping",
-                    context: this.workflowContext,
-                    architectureType: input.architectureType || "nextjs-app-router",
-                },
-                dependencies: ["component-detection"],
-                required: false,
-                retryable: true,
-                interactive: false,
-                // Only run in REUSE_COMPONENTS mode
-                skip: (context) => context.componentDetection?.recommendation !== "REUSE_COMPONENTS",
-            },
-            {
-                id: "component-generation",
-                name: "Component Generation",
-                description: "Generate React components using Claude Agent SDK with MCP integration",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "component-generation",
-                    context: this.workflowContext,
-                    components: "all",
-                },
-                dependencies: ["context-generation"],
-                required: true,
-                retryable: true,
-                interactive: false,
-                // Skip if components should be reused
-                skip: (context) => context.componentDetection?.recommendation === "REUSE_COMPONENTS",
-            },
-            // NEW: Scaffolding Generation Step
-            {
-                id: "scaffolding-generation",
-                name: "Scaffolding Generation",
-                description: "Generate routes, actions, and hooks for components",
-                agentName: "component-first-builder",
-                agent: this,
-                input: {
-                    workflowType: "scaffolding-generation",
-                    context: this.workflowContext,
-                    projectPath: input.projectName,
-                    generateRoutes: input.generateRoutes !== false,
-                    generateActions: input.generateServerActions !== false,
-                    generateHooks: true,
-                    generateLayouts: true,
-                },
-                dependencies: ["component-mapping", "component-generation"],
-                required: false,
-                retryable: true,
-                interactive: false,
-                // Always run if complete architecture is enabled
-                skip: (context) => !input.completeArchitecture,
-            },
-            {
-                id: "validation-testing",
-                name: "Validation & Testing",
-                description: "Validate and test generated components using Claude Agent SDK",
-                agentName: this.name,
-                agent: this,
-                input: {
-                    workflowType: "validation-testing",
-                    context: this.workflowContext,
-                },
-                dependencies: ["component-generation", "scaffolding-generation"],
-                required: true,
-                retryable: true,
-                interactive: false,
-            },
-        ];
-    }
-    /**
-     * Execute enhanced workflow with Claude Agent SDK
-     */
-    async executeEnhancedWorkflow(steps, maxRetries, interactive) {
-        for (const step of steps) {
-            await this.executeStepWithClaudeAgent(step, maxRetries, interactive);
-        }
-    }
-    /**
-     * Execute individual step using Claude Agent SDK
-     */
-    async executeStepWithClaudeAgent(step, maxRetries, interactive) {
-        const stepId = step.id;
-        let retryCount = 0;
-        console.log(chalk_1.default.blue(`\n🔄 Executing: ${step.name}`));
-        console.log(chalk_1.default.gray(`   ${step.description}`));
-        while (retryCount <= maxRetries) {
-            try {
-                // Build step-specific prompt
-                const stepPrompt = this.buildStepPrompt(step);
-                // Execute step using Claude Agent SDK
-                const result = await this.claudeAgentClient.runAgentWorkflow(stepPrompt, this.workflowContext, {
-                    model: "claude-3-5-sonnet-20241022",
-                    temperature: 0.2,
-                    maxTokens: 8000,
-                });
-                // Process step result
-                await this.processStepResult(step, result);
-                // Mark step as completed
-                this.completedSteps.add(stepId);
-                console.log(chalk_1.default.green(`✅ ${step.name} completed successfully`));
-                // Update context with step results
-                this.workflowContext = await this.contextManager.mergeContext(this.workflowContext, {
-                    previousOutputs: {
-                        ...this.workflowContext.previousOutputs,
-                        [stepId]: {
-                            result: result.content,
-                            timestamp: new Date().toISOString(),
-                            success: true,
-                        },
-                    },
-                });
-                return;
-            }
-            catch (error) {
-                retryCount++;
-                this.totalRetries++;
-                if (retryCount <= maxRetries) {
-                    console.log(chalk_1.default.yellow(`⚠️ ${step.name} failed, retrying (${retryCount}/${maxRetries})`));
-                    console.log(chalk_1.default.gray(`   Error: ${error.message}`));
-                    // Add delay before retry
-                    await new Promise((resolve) => setTimeout(resolve, 1000 * retryCount));
-                }
-                else {
-                    console.log(chalk_1.default.red(`❌ ${step.name} failed after ${maxRetries} retries`));
-                    this.failedSteps.push(stepId);
-                    if (step.required) {
-                        throw new Error(`Required step '${step.name}' failed: ${error.message}`);
-                    }
-                }
-            }
-        }
-    }
-    /**
-     * Build step-specific prompt for Claude Agent SDK
-     */
-    buildStepPrompt(step) {
-        const basePrompt = `Execute the following workflow step: ${step.name}\n\nDescription: ${step.description}\n\n`;
-        switch (step.input.workflowType) {
-            case "context-analysis":
-                return basePrompt + this.buildContextAnalysisPrompt(step);
-            case "project-setup":
-                return basePrompt + this.buildProjectSetupPrompt(step);
-            case "requirement-gathering":
-                return basePrompt + this.buildRequirementGatheringPrompt(step);
-            case "context-generation":
-                return basePrompt + this.buildContextGenerationPrompt(step);
-            case "component-generation":
-                return basePrompt + this.buildComponentGenerationPrompt(step);
-            case "validation-testing":
-                return basePrompt + this.buildValidationTestingPrompt(step);
-            default:
-                return (basePrompt +
-                    "Please execute this workflow step according to the provided input.");
-        }
-    }
-    /**
-     * Build context analysis prompt
-     */
-    buildContextAnalysisPrompt(step) {
-        return `Analyze the project context and requirements:
-
-Project Context:
-${JSON.stringify(step.input.projectContext, null, 2)}
-
-Requirements:
-${step.input.requirements}
-
-Please provide:
-1. Context analysis summary
-2. Key requirements identified
-3. Recommended approach
-4. Potential challenges and solutions`;
-    }
-    /**
-     * Build project setup prompt
-     */
-    buildProjectSetupPrompt(step) {
-        return `Set up the project structure:
-
-Project Name: ${step.input.projectName}
-Framework: ${step.input.framework}
-
-Context:
-${JSON.stringify(step.input.context, null, 2)}
-
-Please provide:
-1. Project structure recommendations
-2. Dependencies to install
-3. Configuration files needed
-4. Initial setup steps`;
-    }
-    /**
-     * Build requirement gathering prompt
-     */
-    buildRequirementGatheringPrompt(step) {
-        return `Gather detailed requirements:
-
-Current Context:
-${JSON.stringify(step.input.context, null, 2)}
-
-Interactive Mode: ${step.input.interactive}
-
-Please provide:
-1. Detailed requirements analysis
-2. User stories and use cases
-3. Technical specifications
-4. Acceptance criteria`;
-    }
-    /**
-     * Build context generation prompt
-     */
-    buildContextGenerationPrompt(step) {
-        return `Generate comprehensive context files:
-
-Context:
-${JSON.stringify(step.input.context, null, 2)}
-
-Requirements:
-${step.input.requirements}
-
-Please provide:
-1. PRD (Product Requirements Document)
-2. TypeScript type definitions
-3. Branding guidelines
-4. Component list
-5. Project structure documentation`;
-    }
-    /**
-     * Build component generation prompt
-     */
-    buildComponentGenerationPrompt(step) {
-        return `Generate React components:
-
-Context:
-${JSON.stringify(step.input.context, null, 2)}
-
-Components to generate: ${step.input.components}
-
-Please provide:
-1. Complete React component code
-2. TypeScript interfaces
-3. Styling with Tailwind CSS
-4. Component documentation
-5. Usage examples`;
-    }
-    /**
-     * Build validation testing prompt
-     */
-    buildValidationTestingPrompt(step) {
-        return `Validate and test generated components:
-
-Context:
-${JSON.stringify(step.input.context, null, 2)}
-
-Please provide:
-1. Component validation results
-2. Test cases and scenarios
-3. Performance analysis
-4. Accessibility checks
-5. Recommendations for improvements`;
-    }
-    /**
-     * Process step result
-     */
-    async processStepResult(step, result) {
-        // Process result based on step type
-        switch (step.input.workflowType) {
-            case "context-analysis":
-                await this.processContextAnalysisResult(result);
-                break;
-            case "project-setup":
-                await this.processProjectSetupResult(result);
-                break;
-            case "requirement-gathering":
-                await this.processRequirementGatheringResult(result);
-                break;
-            case "context-generation":
-                await this.processContextGenerationResult(result);
-                break;
-            case "component-generation":
-                await this.processComponentGenerationResult(result);
-                break;
-            case "validation-testing":
-                await this.processValidationTestingResult(result);
-                break;
-        }
-    }
-    /**
-     * Process context analysis result
-     */
-    async processContextAnalysisResult(result) {
-        // Extract key insights from context analysis
-        console.log(chalk_1.default.green("📊 Context analysis completed"));
-    }
-    /**
-     * Process project setup result
-     */
-    async processProjectSetupResult(result) {
-        // Process project setup recommendations
-        console.log(chalk_1.default.green("🏗️ Project setup completed"));
-    }
-    /**
-     * Process requirement gathering result
-     */
-    async processRequirementGatheringResult(result) {
-        // Process gathered requirements
-        console.log(chalk_1.default.green("📋 Requirements gathered"));
-    }
-    /**
-     * Process context generation result
-     */
-    async processContextGenerationResult(result) {
-        // Process generated context files
-        console.log(chalk_1.default.green("📄 Context files generated"));
-    }
-    /**
-     * Process component generation result
-     */
-    async processComponentGenerationResult(result) {
-        // Process generated components
-        console.log(chalk_1.default.green("🎨 Components generated"));
-    }
-    /**
-     * Process validation testing result
-     */
-    async processValidationTestingResult(result) {
-        // Process validation results
-        console.log(chalk_1.default.green("✅ Validation completed"));
-    }
-    /**
-     * Get enhanced system prompt for Claude Agent SDK
-     */
-    getEnhancedSystemPrompt() {
-        return `You are an expert AI assistant specialized in React/Next.js development and component generation, enhanced with Claude Agent SDK capabilities.
-
-Your enhanced capabilities include:
-- Advanced context management and compaction
-- Fine-grained tool permissions
-- MCP (Model Context Protocol) integration
-- Intelligent workflow orchestration
-- Automatic error handling and retry logic
-
-When executing workflows:
-1. Use context management to maintain project state
-2. Respect tool permissions and restrictions
-3. Leverage MCP tools when available
-4. Provide detailed progress updates
-5. Handle errors gracefully with retry logic
-6. Maintain comprehensive project context
-
-Provide intelligent, context-aware responses that demonstrate deep understanding of the project and requirements.`;
-    }
-}
-exports.ClaudeAgentWorkflow = ClaudeAgentWorkflow;
-//# sourceMappingURL=ClaudeAgentWorkflow.js.map