Hi-AI

Instructions

Implementation Reference

• src/tools/prompt/analyzePrompt.ts:30-182 (handler)

  The async handler function that analyzes the prompt based on criteria like clarity, specificity, context, and structure. Computes scores, feedback, strengths, weaknesses, and recommendations, returning formatted ToolResult.
  export async function analyzePrompt(args: { prompt: string; criteria?: string[] }): Promise<ToolResult> { const { prompt, criteria = ['clarity', 'specificity', 'context', 'structure'] } = args; // Initialize scores with explicit types const scores: Record<string, number> = {}; const feedback: Record<string, string[]> = {}; // Analyze clarity (0-10) if (criteria.includes('clarity')) { let clarityScore = 5.0; const clarityFeedback: string[] = []; if (prompt.length < 20) { clarityScore -= 2.0; clarityFeedback.push('프롬프트가 너무 짧습니다'); } if (prompt.includes('?') || /\b(해주세요|부탁|요청)\b/.test(prompt)) { clarityScore += 2.0; clarityFeedback.push('명확한 요청 형태 ✓'); } if (prompt.split(',').length > 5 || prompt.split('.').length > 10) { clarityScore -= 1.0; clarityFeedback.push('문장이 너무 복잡합니다'); } scores.clarity = Math.max(0, Math.min(10, clarityScore)); feedback.clarity = clarityFeedback; } // Analyze specificity (0-10) if (criteria.includes('specificity')) { let specificityScore = 5.0; const specificityFeedback: string[] = []; const specificKeywords = ['구체적', '정확히', '예시', '예를 들어']; const hasSpecificWords = specificKeywords.some(word => prompt.includes(word)); if (hasSpecificWords) { specificityScore += 2.0; specificityFeedback.push('구체적인 표현 사용 ✓'); } const techTerms = /\b(JavaScript|Python|React|Node\.js|API|데이터베이스)\b/i; if (techTerms.test(prompt)) { specificityScore += 2.0; specificityFeedback.push('기술 용어 포함 ✓'); } if (!prompt.match(/\d+/) && prompt.length > 50) { specificityScore -= 1.0; specificityFeedback.push('수치나 구체적 데이터 부족'); } scores.specificity = Math.max(0, Math.min(10, specificityScore)); feedback.specificity = specificityFeedback; } // Analyze context (0-10) if (criteria.includes('context')) { let contextScore = 5.0; const contextFeedback: string[] = []; const contextKeywords = ['배경', '목적', '이유', '상황', '현재', '문제']; const contextCount = contextKeywords.filter(word => prompt.includes(word)).length; contextScore += contextCount * 1.5; if (contextCount > 0) { contextFeedback.push(`배경 정보 포함 (${contextCount}개 키워드) ✓`); } else { contextFeedback.push('배경 정보 부족'); } if (prompt.split('\n').length > 2) { contextScore += 1.0; contextFeedback.push('구조화된 설명 ✓'); } scores.context = Math.max(0, Math.min(10, contextScore)); feedback.context = contextFeedback; } // Analyze structure (0-10) if (criteria.includes('structure')) { let structureScore = 5.0; const structureFeedback: string[] = []; if (prompt.includes('\n')) { structureScore += 2.0; structureFeedback.push('줄바꿈 사용 ✓'); } if (/[1-9]\.|[-•]/.test(prompt)) { structureScore += 2.0; structureFeedback.push('목록 형식 사용 ✓'); } if (prompt.includes('**') || prompt.includes('##')) { structureScore += 1.0; structureFeedback.push('마크다운 형식 사용 ✓'); } scores.structure = Math.max(0, Math.min(10, structureScore)); feedback.structure = structureFeedback; } // Calculate total score const totalScore = Object.values(scores).reduce((a, b) => a + b, 0) / Object.keys(scores).length; // Generate recommendations const recommendations: string[] = []; if (scores.clarity < 6) { recommendations.push('💡 질문이나 요청을 더 명확하게 표현하세요'); } if (scores.specificity < 6) { recommendations.push('💡 구체적인 예시나 기술 사양을 추가하세요'); } if (scores.context < 6) { recommendations.push('💡 작업의 배경과 목적을 설명하세요'); } if (scores.structure < 6) { recommendations.push('💡 번호나 불릿 포인트로 구조화하세요'); } // Identify strengths and weaknesses const strengths = Object.entries(scores) .filter(([_, score]) => score >= 7) .map(([category, score]) => `✨ ${category}: 우수함 (${score.toFixed(1)}/10)`); const weaknesses = Object.entries(scores) .filter(([_, score]) => score < 5) .map(([category, score]) => `⚠️ ${category}: 개선 필요 (${score.toFixed(1)}/10)`); const analysis = { action: 'analyze_prompt', prompt, totalScore: parseFloat(totalScore.toFixed(1)), scores: Object.fromEntries( Object.entries(scores).map(([k, v]) => [k, parseFloat(v.toFixed(1))]) ), feedback, strengths, weaknesses, recommendations, grade: totalScore >= 8 ? 'A' : totalScore >= 6 ? 'B' : totalScore >= 4 ? 'C' : 'D', status: 'success' }; return { content: [{ type: 'text', text: `Score: ${analysis.totalScore}/10 (Grade: ${analysis.grade})\n\nScores:\n${Object.entries(analysis.scores).map(([k, v]) => `- ${k}: ${v}/10`).join('\n')}\n\nStrengths:\n${analysis.strengths.length > 0 ? analysis.strengths.join('\n') : 'None identified'}\n\nWeaknesses:\n${analysis.weaknesses.length > 0 ? analysis.weaknesses.join('\n') : 'None identified'}\n\nRecommendations:\n${analysis.recommendations.map(r => `- ${r}`).join('\n')}` }] }; }
• src/tools/prompt/analyzePrompt.ts:5-28 (schema)

  The ToolDefinition object defining the input schema, description, and annotations for the analyze_prompt tool.
  export const analyzePromptDefinition: ToolDefinition = { name: 'analyze_prompt', description: '프롬프트 분석|평가|점수|얼마나 좋은지|analyze prompt|rate this|score|how good|prompt quality - Analyze prompt quality', inputSchema: { type: 'object', properties: { prompt: { type: 'string', description: 'Prompt to analyze' }, criteria: { type: 'array', items: { type: 'string' }, description: 'Specific criteria to evaluate (default: all)' } }, required: ['prompt'] }, annotations: { title: 'Analyze Prompt', audience: ['user', 'assistant'], readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false } };
• src/index.ts:137-137 (registration)

  Registration of the tool definition in the tools array used for ListToolsRequest.
  analyzePromptDefinition,
• src/index.ts:204-204 (registration)

  Registration of the tool handler in the toolHandlers map for dynamic dispatch during tool execution.
  'analyze_prompt': analyzePrompt,