diff --git a/.claude/skills/prompt-enhancer/SKILL.md b/.claude/skills/prompt-enhancer/SKILL.md index d3dfa36d..6afec116 100644 --- a/.claude/skills/prompt-enhancer/SKILL.md +++ b/.claude/skills/prompt-enhancer/SKILL.md @@ -1,12 +1,12 @@ --- name: Prompt Enhancer -description: Transform vague prompts into actionable specs using session memory + semantic analysis. AUTO-TRIGGER on (1) -e/--enhance flags, (2) vague keywords (fix/improve/refactor/修复/优化/重构), (3) unclear refs (it/that/这个/那个), (4) multi-module scope. Supports English + Chinese semantic recognition. -allowed-tools: Bash, Read, Glob, Grep, AskUserQuestion +description: Transform vague prompts into actionable specs using session memory ONLY (no file analysis). AUTO-TRIGGER on (1) -e/--enhance flags, (2) vague keywords (fix/improve/refactor/修复/优化/重构), (3) unclear refs (it/that/这个/那个), (4) multi-module scope. Supports English + Chinese semantic recognition. +allowed-tools: AskUserQuestion --- # Prompt Enhancer -**Transform**: Vague intent → Structured specification +**Transform**: Vague intent → Structured specification (Memory-based) **Languages**: English + Chinese (中英文语义识别) @@ -19,33 +19,47 @@ allowed-tools: Bash, Read, Glob, Grep, AskUserQuestion | **P3** | Unclear references | EN: it/that/the code
CN: 这个/那个/它/代码 | Context extraction | | **P4** | Multi-module scope | >3 modules or critical systems | Dependency analysis | -## Process (4 Steps) +## Process (3 Steps - Streamlined) -### 1. Semantic Analysis +### 1. Semantic Analysis (Quick) Identify: Intent keywords (EN/CN) → Scope (file/module/system) → Domain (auth/API/DB/UI) **English**: fix, improve, add, refactor, update, migrate **Chinese**: 修复, 优化, 添加, 重构, 更新, 迁移, 改进, 清理 -### 2. Memory Extraction -Extract: Technical context → Current state → Dependencies → Constraints +### 2. Memory-Only Extraction (NO File Reading) +Extract from **conversation memory ONLY**: +- Recent user requests and context +- Previous implementations/patterns discussed +- Known dependencies from session +- User preferences and constraints -### 3. Integration -Combine: Precise intent + Action steps (with files) + Critical constraints +**SKIP**: File reading, codebase scanning, Glob/Grep operations +**FOCUS**: Pure memory-based context extraction -### 4. User Confirmation (REQUIRED) -Use AskUserQuestion with 3 options → Execute/Review/Refine +### 3. User Confirmation with Optimization Direction (REQUIRED) +Present structured prompt → Ask: Continue? + Optimization suggestions needed? ## Output Format ``` -INTENT: [One-sentence technical goal] -CONTEXT: [Session memory + domain analysis] +📋 ENHANCED PROMPT + +INTENT: [One-sentence technical goal / 明确技术目标] + +CONTEXT: [Session memory findings / 会话记忆发现] +- [Key context point 1] +- [Key context point 2] +- [...] + ACTION: -1. [Step with file references] -2. [Step with file references] +1. [Concrete step / 具体步骤] +2. [Concrete step / 具体步骤] 3. [...] -ATTENTION: [Security/compatibility/testing constraints] + +ATTENTION: [Critical constraints / 关键约束] +- [Security/compatibility/testing concerns] + ``` ## Semantic Patterns (EN + CN) @@ -59,48 +73,50 @@ ATTENTION: [Security/compatibility/testing constraints] | update/更新 + version | Modernize | Version compatibility | | clean up/清理 + area | Simplify/organize | Remove redundancy | -### Scope Detection - -| Input Pattern | Scope | Action | -|---------------|-------|--------| -| "fix button", "修复按钮" | Single file | Session memory only | -| "add auth", "添加认证" | Multi-module (>3) | Dependency analysis | -| "improve perf", "优化性能" | System-wide | Cross-cutting concerns | - ## Workflow ``` -Trigger → Analyze → Enhance → Confirm → Execute - ↓ ↓ ↓ ↓ ↓ - P1-4 EN/CN INTENT/ 3 opts Direct - detect ACTION or iterate +Trigger → Analyze → Extract → Present → Confirm → Execute + ↓ ↓ ↓ ↓ ↓ ↓ + P1-4 EN/CN Memory Struct Ask User Direct + detect only prompt or refine ``` 1. **Detect**: Check triggers (P1-P4) -2. **Analyze**: Semantic (EN/CN) + Memory extraction -3. **Enhance**: Generate structured output -4. **Confirm**: AskUserQuestion (Execute/Review/Refine) -5. **Execute**: Proceed based on user choice +2. **Analyze**: Semantic (EN/CN) analysis of user intent +3. **Extract**: Memory-only context extraction (NO file reading) +4. **Present**: Generate structured prompt output +5. **Confirm**: AskUserQuestion (Continue/Modify/Cancel) +6. **Execute**: Proceed based on user choice ## Confirmation (AskUserQuestion) -**Question**: "已增强提示词，如何继续？/ Enhanced prompt ready. Proceed?" +**Question**: "Enhanced prompt ready. Proceed or need adjustments? (已生成增强提示词，是否继续或需要调整？)" **Options**: -1. **Execute now** / **立即执行** - Proceed immediately -2. **Review first** / **先查看** - Show full spec before execution -3. **Refine** / **优化需求** - Iterate with user input +1. **Continue as-is / 按此继续** - Proceed with current specification +2. **Suggest optimizations / 建议优化方向** - I need guidance on how to improve this +3. **Modify requirements / 修改需求** - Let me provide specific changes ```typescript AskUserQuestion({ questions: [{ - question: "I've enhanced your prompt. How would you like to proceed? (已增强提示词，如何继续？)", - header: "Execution", + question: "Enhanced prompt ready. Proceed or need adjustments? (已生成增强提示词，是否继续或需要调整？)", + header: "Next Step", multiSelect: false, options: [ - { label: "Execute now", description: "Proceed immediately with enhanced spec" }, - { label: "Review first", description: "Show full specification before execution" }, - { label: "Refine", description: "Let me adjust the requirements" } + { + label: "Continue as-is", + description: "Proceed with current specification (按此继续)" + }, + { + label: "Suggest optimizations", + description: "I need guidance on how to improve this (建议优化方向)" + }, + { + label: "Modify requirements", + description: "Let me provide specific changes (修改需求)" + } ] }] }) @@ -110,23 +126,53 @@ AskUserQuestion({ - ✅ Detect `-e`/`--enhance` flags first (P1) - ✅ Support EN + CN semantic keywords -- ✅ Extract session memory context +- ✅ Extract **memory context ONLY** (no file reading) - ✅ Use INTENT/CONTEXT/ACTION/ATTENTION format - ✅ ALWAYS confirm with AskUserQuestion -- ✅ Execute directly if "Execute now" selected -- ✅ Include file references in ACTION steps +- ✅ Offer optimization guidance option +- ❌ NO Bash, Read, Glob, Grep operations +- ❌ NO direct file analysis + +## Key Changes from Previous Version + +1. **Removed file analysis** - Memory extraction only +2. **Simplified to 3 steps** - Faster workflow +3. **Updated confirmation options** - Added "Suggest optimizations" +4. **Removed file tools** - Only AskUserQuestion allowed +5. **Focus on speed** - Quick semantic analysis + memory extraction ## Examples **Input**: "fix auth -e" / "优化性能 --enhance" +**Process**: +1. Detect P1 trigger (`-e` flag) +2. Semantic analysis: "fix/优化" intent +3. Extract from memory: Recent auth/performance discussions +4. Generate structured prompt +5. Ask user: Continue/Suggest optimizations/Modify + **Output**: ``` -INTENT: [Clear technical goal / 明确技术目标] -CONTEXT: [Session memory findings / 会话记忆发现] +📋 ENHANCED PROMPT + +INTENT: Fix authentication module issues based on recent session context + +CONTEXT: +- User mentioned auth token expiration problems +- Previous discussion about JWT validation +- Session indicates preference for backward compatibility + ACTION: -1. [Step with files / 带文件引用的步骤] -2. [Step with files / 带文件引用的步骤] -ATTENTION: [Constraints / 约束条件] +1. Review authentication token handling logic +2. Implement proper JWT validation with expiration checks +3. Add unit tests for token refresh flow +4. Update documentation for auth changes + +ATTENTION: +- Must maintain backward compatibility with existing tokens +- Security: Follow JWT best practices +- Testing: Ensure no breaking changes to API contracts ``` -**Confirm** → Execute/Review/Refine + +**Then ask**: Continue as-is / Suggest optimizations / Modify requirements