AI Code Toolkit

/** * Review Code Change Command * * DESIGN PATTERNS: * - Command pattern with Commander for CLI argument parsing * - Async/await pattern for asynchronous operations * - Error handling pattern with try-catch and proper exit codes * * CODING STANDARDS: * - Use async action handlers for asynchronous operations * - Provide clear option descriptions and default values * - Handle errors gracefully with process.exit() * - Log progress and errors to console * - Use Commander's .option() and .argument() for inputs * * AVOID: * - Synchronous blocking operations in action handlers * - Missing error handling (always use try-catch) * - Hardcoded values (use options or environment variables) * - Not exiting with appropriate exit codes on errors */ import { log, print } from '@agiflowai/aicode-utils'; import { Command } from 'commander'; import { ReviewCodeChangeTool } from '../tools/ReviewCodeChangeTool'; import { type LlmToolId, isValidLlmTool, SUPPORTED_LLM_TOOLS, } from '@agiflowai/coding-agent-bridge'; interface ReviewCodeChangeOptions { verbose?: boolean; json?: boolean; llmTool?: string; toolConfig?: string; } /** * Review code changes against template-specific and global rules to identify violations */ export const reviewCodeChangeCommand = new Command('review-code-change') .description( 'Review code changes against template-specific and global rules to identify violations', ) .argument('<file-path>', 'Path to the file to review') .option('-v, --verbose', 'Enable verbose output', false) .option('-j, --json', 'Output as JSON', false) .option( '--llm-tool <tool>', `LLM tool to use for code review. Supported: ${SUPPORTED_LLM_TOOLS.join(', ')}`, 'claude-code', ) .option( '--tool-config <json>', 'JSON config for the LLM tool (e.g., \'{"model":"gpt-5.2-high"}\')', undefined, ) .action(async (filePath: string, options: ReviewCodeChangeOptions): Promise<void> => { try { if (options.verbose) { print.info(`Reviewing file: ${filePath}`); print.info(`Using LLM tool: ${options.llmTool}`); } // Validate llm-tool option if (options.llmTool && !isValidLlmTool(options.llmTool)) { print.error( `❌ Error: Invalid LLM tool "${options.llmTool}". Supported: ${SUPPORTED_LLM_TOOLS.join(', ')}`, ); process.exit(1); } // Parse tool config JSON let toolConfig: Record<string, unknown> | undefined; if (options.toolConfig) { try { toolConfig = JSON.parse(options.toolConfig); if (options.verbose) { print.info(`Tool config: ${JSON.stringify(toolConfig)}`); } } catch (error) { print.error( `❌ Error: Invalid JSON for --tool-config: ${error instanceof Error ? error.message : String(error)}`, ); process.exit(1); } } // Create tool instance const tool = new ReviewCodeChangeTool({ llmTool: options.llmTool as LlmToolId | undefined, toolConfig, }); // Execute the tool const result = await tool.execute({ file_path: filePath, }); // Parse and display result const firstContent = result.content[0]; if (firstContent.type !== 'text') { print.error('❌ Error: Unexpected response type'); process.exit(1); } if (result.isError) { const errorData = JSON.parse(firstContent.text); print.error('❌ Error:', errorData.error || errorData); process.exit(1); } const data = JSON.parse(firstContent.text); if (options.json) { // Output raw JSON print.info(JSON.stringify(data, null, 2)); } else { // Pretty print the results print.info(`\n## ${data.file_path}`); if (data.project_name) { print.info(`Project: ${data.project_name}`); } if (data.source_template) { print.info(`Template: ${data.source_template}`); } if (data.matched_rules) { print.info('\n### Matched Rule Pattern\n'); print.info(`**${data.matched_rules.pattern}**`); print.info(data.matched_rules.description); print.info(''); } // Display fix required status const fixRequiredStatus = data.fix_required ? '🔴 Fix Required' : '🟢 No Fixes Required'; print.info(`### Review Result: ${fixRequiredStatus}\n`); // Display feedback if (data.review_feedback) { print.info(data.review_feedback); print.info(''); } // Display issues found if (data.issues_found && data.issues_found.length > 0) { print.info('### Issues Found\n'); const groupedIssues: Record<string, any[]> = { must_not_do: [], must_do: [], should_do: [], }; for (const issue of data.issues_found) { if (groupedIssues[issue.type]) { groupedIssues[issue.type].push(issue); } } if (groupedIssues.must_not_do.length > 0) { print.info('**❌ Must Not Do Violations:**\n'); for (const issue of groupedIssues.must_not_do) { print.info(`- ${issue.rule}`); if (issue.violation) { print.info(` Violation: ${issue.violation}`); } } print.info(''); } if (groupedIssues.must_do.length > 0) { print.info('**⚠️ Must Do Missing:**\n'); for (const issue of groupedIssues.must_do) { print.info(`- ${issue.rule}`); if (issue.violation) { print.info(` Note: ${issue.violation}`); } } print.info(''); } if (groupedIssues.should_do.length > 0) { print.info('**💡 Should Do Suggestions:**\n'); for (const issue of groupedIssues.should_do) { print.info(`- ${issue.rule}`); if (issue.violation) { print.info(` Note: ${issue.violation}`); } } print.info(''); } } else { print.success('✅ No violations found!\n'); } } } catch (error) { log.error('❌ Error executing review-code-change:', error); process.exit(1); } });