@chongyan/autospec 1.0.1 → 1.0.2

package/plugins/benchmarks/templates/skills/implementation-stage-template.yaml

@@ -0,0 +1,127 @@
+# AutoSpec Skill Benchmark Template - Implementation-Stage
+# 适用于: 测试 implementation-stage skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-implementation-stage"
+description: "Implementation-Stage Skill 基准测试"
+
+type: skill
+target: implementation-stage
+
+testCases:
+  - name: "simple-implementation"
+    input:
+      context: "实现 {feature-name} API"
+    complexity: 1
+    expectedBehaviors:
+      - "读取设计文档"
+      - "实现 API 端点"
+      - "实现业务逻辑"
+      - "编写单元测试"
+      - "执行 Layer1 验证"
+    expectedOutput:
+      - "src/{feature}/*.js"
+      - "tests/{feature}/*.test.js"
+      - "Layer1 验证报告"
+    successCriteria:
+      - "API 实现符合设计文档"
+      - "单元测试覆盖核心逻辑"
+      - "Layer1 验证通过"
+    qualityMetrics:
+      - "测试覆盖率 >= 70%"
+      - "代码质量评分 >= 80%"
+    maxDuration: 600
+
+  - name: "multi-system-implementation"
+    input:
+      context: "实现 {feature}，包含后端和前端"
+    complexity: 5
+    expectedBehaviors:
+      - "读取多系统设计文档"
+      - "先实现后端 API"
+      - "执行后端 Layer1 验证"
+      - "再实现前端页面"
+      - "执行集成验证"
+    expectedOutput:
+      - "backend/src/{feature}/*.js"
+      - "frontend/src/pages/{feature}/*.jsx"
+      - "集成测试报告"
+    successCriteria:
+      - "依赖顺序正确"
+      - "集成测试通过"
+    qualityMetrics:
+      - "后端测试覆盖率 >= 70%"
+      - "集成测试通过率 >= 90%"
+    maxDuration: 1200
+
+  - name: "brownfield-implementation"
+    input:
+      context: "在现有项目中添加 {feature} 功能"
+    complexity: 3
+    expectedBehaviors:
+      - "执行 Brownfield 扫描"
+      - "识别现有代码结构"
+      - "复用现有依赖包"
+      - "遵循现有代码风格"
+    expectedOutput:
+      - "现有代码分析报告"
+      - "src/{feature}/*.js"
+      - "依赖复用检查报告"
+    successCriteria:
+      - "正确识别现有代码结构"
+      - "遵循现有代码风格"
+      - "与现有代码无缝集成"
+    qualityMetrics:
+      - "Brownfield 扫描执行率 = 100%"
+      - "代码风格符合率 >= 90%"
+    maxDuration: 900
+
+  - name: "ai-implementation"
+    input:
+      context: "实现 LLM 智能问答功能"
+    complexity: 5
+    expectedBehaviors:
+      - "实现模型加载逻辑"
+      - "实现推理接口"
+      - "实现效果评测脚本"
+      - "执行效果评测验证"
+    expectedOutput:
+      - "ai/model_loader.py"
+      - "ai/inference.py"
+      - "ai/evaluation.py"
+      - "效果评测报告"
+    successCriteria:
+      - "模型加载正确"
+      - "推理接口响应时间 < 2s"
+      - "评测指标达标"
+    qualityMetrics:
+      - "AI 组件完整率 >= 90%"
+      - "响应时间达标率 = 100%"
+    maxDuration: 900
+
+  - name: "db-migration-implementation"
+    input:
+      context: "实现数据库迁移，添加 {field} 字段"
+    complexity: 3
+    expectedBehaviors:
+      - "读取数据库设计文档"
+      - "编写迁移脚本（向上/向下）"
+      - "验证迁移安全性"
+      - "执行迁移测试"
+    expectedOutput:
+      - "migrations/add_{field}.js"
+      - "migrations/add_{field}.down.js"
+      - "迁移验证报告"
+    successCriteria:
+      - "迁移脚本完整"
+      - "迁移安全性验证通过"
+      - "数据完整性检查通过"
+    qualityMetrics:
+      - "迁移安全性 = 100%"
+      - "数据完整性 = 100%"
+    maxDuration: 450
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/layer1-validation-template.yaml

@@ -0,0 +1,121 @@
+# AutoSpec Skill Benchmark Template - Layer1-Validation
+# 适用于: 测试 layer1-validation skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-layer1-validation"
+description: "Layer1-Validation Skill 基准测试"
+
+type: skill
+target: layer1-validation
+
+testCases:
+  - name: "single-system-validation"
+    input:
+      context: "对 Node.js 项目进行 Layer1 验证"
+    complexity: 1
+    expectedBehaviors:
+      - "识别项目类型"
+      - "执行编译检查"
+      - "执行测试"
+      - "执行 Lint 检查"
+      - "执行类型检查"
+    expectedOutput:
+      - "验证报告表格"
+      - "各项检查状态（PASS/FAIL/BLOCKED）"
+      - "日志摘要"
+    successCriteria:
+      - "验证步骤完整执行"
+      - "结果真实（有执行日志）"
+      - "状态标注正确"
+    qualityMetrics:
+      - "验证执行率 = 100%"
+      - "结果准确率 = 100%"
+    maxDuration: 300
+
+  - name: "multi-system-validation"
+    input:
+      context: "对多系统项目（backend/frontend）进行 Layer1 验证"
+    complexity: 3
+    expectedBehaviors:
+      - "读取 config.json 获取 subsystems 配置"
+      - "按依赖顺序验证"
+      - "对每个子系统执行 build/test/lint/typeCheck"
+      - "汇总验证结果"
+    expectedOutput:
+      - "多系统验证报告"
+      - "各子系统验证状态"
+      - "整体验证结论"
+    successCriteria:
+      - "子系统识别正确"
+      - "验证顺序正确"
+      - "所有子系统验证完成"
+    qualityMetrics:
+      - "子系统识别率 = 100%"
+      - "验证顺序正确率 = 100%"
+    maxDuration: 600
+
+  - name: "ai-subsystem-validation"
+    input:
+      context: "对 AI 子系统进行 Layer1 验证"
+    complexity: 3
+    expectedBehaviors:
+      - "识别 AI 子系统类型"
+      - "执行模型加载验证"
+      - "执行推理验证"
+      - "执行评测脚本验证"
+    expectedOutput:
+      - "AI 子系统验证报告"
+      - "模型加载状态"
+      - "推理验证结果"
+      - "评测脚本执行结果"
+    successCriteria:
+      - "AI 组件识别正确"
+      - "评测脚本执行完整"
+    qualityMetrics:
+      - "AI 组件识别率 = 100%"
+      - "评测执行率 = 100%"
+    maxDuration: 600
+
+  - name: "data-subsystem-validation"
+    input:
+      context: "对数据子系统进行 Layer1 验证"
+    complexity: 3
+    expectedBehaviors:
+      - "识别数据子系统类型"
+      - "执行数据质量检查"
+      - "执行 ETL 流程验证"
+    expectedOutput:
+      - "数据子系统验证报告"
+      - "数据质量检查结果"
+      - "ETL 流程验证结果"
+    successCriteria:
+      - "数据组件识别正确"
+      - "数据质量检查完整"
+    qualityMetrics:
+      - "数据组件识别率 = 100%"
+      - "数据质量检查率 = 100%"
+    maxDuration: 450
+
+  - name: "validation-failure"
+    input:
+      context: "Layer1 验证失败时的正确处理"
+    complexity: 1
+    expectedBehaviors:
+      - "执行验证"
+      - "检测失败项"
+      - "状态标注为 FAIL"
+      - "记录失败日志"
+      - "不进入 Layer2 审查"
+    successCriteria:
+      - "失败正确识别"
+      - "状态标注正确"
+      - "正确阻止进入 Layer2"
+    qualityMetrics:
+      - "失败检出率 = 100%"
+      - "状态标注准确率 = 100%"
+    maxDuration: 180
+
+successCriteria:
+  passRate: 90
+  avgFieldCompletion: 95

package/plugins/benchmarks/templates/skills/project-evolve-analyzer-template.yaml

@@ -0,0 +1,51 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Analyzer
+# 适用于: 测试 project-evolve-analyzer skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-analyzer"
+description: "Project-Evolve-Analyzer Skill 基准测试"
+
+type: skill
+target: project-evolve-analyzer
+
+testCases:
+  - name: "problem-analysis"
+    input:
+      context: "分析项目问题"
+    complexity: 3
+    expectedBehaviors:
+      - "读取扫描结果"
+      - "识别问题模式"
+      - "分析根因"
+      - "评估影响范围"
+    expectedOutput:
+      - "问题分析报告"
+      - "根因分析"
+      - "影响评估"
+    successCriteria:
+      - "问题模式识别准确"
+      - "根因分析深入"
+    qualityMetrics:
+      - "根因识别率 >= 85%"
+    maxDuration: 300
+
+  - name: "priority-calculation"
+    input:
+      context: "计算问题修复优先级"
+    complexity: 1
+    expectedBehaviors:
+      - "评估影响程度"
+      - "评估修复难度"
+      - "计算优先级分数"
+    expectedOutput:
+      - "优先级排序列表"
+    successCriteria:
+      - "优先级计算合理"
+    qualityMetrics:
+      - "优先级合理性 >= 90%"
+    maxDuration: 120
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-fixer-template.yaml

@@ -0,0 +1,52 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Fixer
+# 适用于: 测试 project-evolve-fixer skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-fixer"
+description: "Project-Evolve-Fixer Skill 基准测试"
+
+type: skill
+target: project-evolve-fixer
+
+testCases:
+  - name: "auto-fix"
+    input:
+      context: "自动修复检测到的问题"
+    complexity: 3
+    expectedBehaviors:
+      - "读取问题清单"
+      - "分类可自动修复问题"
+      - "执行修复"
+      - "验证修复效果"
+    expectedOutput:
+      - "修复记录"
+      - "验证报告"
+    successCriteria:
+      - "修复有效"
+      - "无退化发生"
+    qualityMetrics:
+      - "修复有效率 >= 85%"
+      - "退化检出率 = 100%"
+    maxDuration: 600
+
+  - name: "manual-review-required"
+    input:
+      context: "处理需要人工审查的问题"
+    complexity: 1
+    expectedBehaviors:
+      - "识别需人工审查问题"
+      - "生成审查建议"
+      - "不自动执行"
+    expectedOutput:
+      - "审查建议清单"
+    successCriteria:
+      - "分类正确"
+      - "不自动执行"
+    qualityMetrics:
+      - "分类准确率 = 100%"
+    maxDuration: 120
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-generator-template.yaml

@@ -0,0 +1,34 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Generator
+# 适用于: 测试 project-evolve-generator skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-generator"
+description: "Project-Evolve-Generator Skill 基准测试"
+
+type: skill
+target: project-evolve-generator
+
+testCases:
+  - name: "generate-benchmark"
+    input:
+      context: "生成项目 benchmark"
+    complexity: 3
+    expectedBehaviors:
+      - "检测项目类型"
+      - "生成基础 benchmark"
+      - "生成 pipeline benchmark"
+    expectedOutput:
+      - "project-generated.yaml"
+      - "generate-report.md"
+    successCriteria:
+      - "项目类型识别正确"
+      - "benchmark 结构完整"
+    qualityMetrics:
+      - "类型识别率 = 100%"
+      - "benchmark 完整率 >= 90%"
+    maxDuration: 300
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-learner-template.yaml

@@ -0,0 +1,50 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Learner
+# 适用于: 测试 project-evolve-learner skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-learner"
+description: "Project-Evolve-Learner Skill 基准测试"
+
+type: skill
+target: project-evolve-learner
+
+testCases:
+  - name: "pattern-learning"
+    input:
+      context: "从实践日志中学习模式"
+    complexity: 3
+    expectedBehaviors:
+      - "读取 practice-log"
+      - "识别重复模式"
+      - "提取通用解决方案"
+    expectedOutput:
+      - "模式识别报告"
+      - "通用解决方案"
+    successCriteria:
+      - "模式识别准确"
+      - "解决方案合理"
+    qualityMetrics:
+      - "模式识别率 >= 80%"
+    maxDuration: 300
+
+  - name: "knowledge-update"
+    input:
+      context: "更新项目知识"
+    complexity: 3
+    expectedBehaviors:
+      - "分析新发现"
+      - "更新知识文件"
+      - "验证更新效果"
+    expectedOutput:
+      - "更新后的知识文件"
+      - "更新报告"
+    successCriteria:
+      - "知识更新正确"
+    qualityMetrics:
+      - "更新准确率 >= 90%"
+    maxDuration: 300
+
+successCriteria:
+  passRate: 80
+  avgFieldCompletion: 85

package/plugins/benchmarks/templates/skills/project-evolve-reviewer-template.yaml

@@ -0,0 +1,50 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Reviewer
+# 适用于: 测试 project-evolve-reviewer skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-reviewer"
+description: "Project-Evolve-Reviewer Skill 基准测试"
+
+type: skill
+target: project-evolve-reviewer
+
+testCases:
+  - name: "code-review"
+    input:
+      context: "审查代码变更"
+    complexity: 3
+    expectedBehaviors:
+      - "读取变更内容"
+      - "执行代码审查"
+      - "识别问题"
+      - "生成审查报告"
+    expectedOutput:
+      - "审查报告"
+      - "问题清单"
+    successCriteria:
+      - "审查维度完整"
+      - "问题识别准确"
+    qualityMetrics:
+      - "问题检出率 >= 85%"
+    maxDuration: 300
+
+  - name: "design-review"
+    input:
+      context: "审查设计文档"
+    complexity: 3
+    expectedBehaviors:
+      - "读取设计文档"
+      - "执行设计审查"
+      - "检查一致性"
+    expectedOutput:
+      - "设计审查报告"
+    successCriteria:
+      - "设计审查完整"
+    qualityMetrics:
+      - "一致性检查执行率 = 100%"
+    maxDuration: 300
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-scanner-template.yaml

@@ -0,0 +1,75 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Scanner
+# 适用于: 测试 project-evolve-scanner skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-scanner"
+description: "Project-Evolve-Scanner Skill 基准测试"
+
+type: skill
+target: project-evolve-scanner
+
+testCases:
+  - name: "nodejs-project-scan"
+    input:
+      context: "扫描一个 Node.js + React 项目"
+    complexity: 1
+    expectedBehaviors:
+      - "检测 package.json"
+      - "识别技术栈"
+      - "分析目录结构"
+      - "发现评测点"
+    expectedOutput:
+      - "项目特征报告"
+      - "技术栈列表"
+      - "评测点列表"
+    successCriteria:
+      - "技术栈检测正确"
+      - "评测点发现完整"
+    qualityMetrics:
+      - "技术栈识别准确率 >= 95%"
+      - "评测点发现率 >= 80%"
+    maxDuration: 300
+
+  - name: "java-project-scan"
+    input:
+      context: "扫描一个 Java Spring Boot 项目"
+    complexity: 3
+    expectedBehaviors:
+      - "检测 pom.xml 或 build.gradle"
+      - "识别技术栈"
+      - "分析依赖版本"
+    expectedOutput:
+      - "项目特征报告"
+      - "依赖分析报告"
+    successCriteria:
+      - "技术栈检测正确"
+      - "依赖分析完整"
+    qualityMetrics:
+      - "技术栈识别准确率 >= 95%"
+      - "依赖分析完整率 >= 90%"
+    maxDuration: 450
+
+  - name: "multi-system-scan"
+    input:
+      context: "扫描包含前后端的多子系统项目"
+    complexity: 5
+    expectedBehaviors:
+      - "检测多个子系统"
+      - "识别各子系统技术栈"
+      - "分析子系统依赖关系"
+      - "生成记忆结构"
+    expectedOutput:
+      - "多系统特征报告"
+      - "系统间依赖关系图"
+    successCriteria:
+      - "子系统识别完整"
+      - "依赖关系清晰"
+    qualityMetrics:
+      - "子系统识别率 = 100%"
+      - "技术栈识别准确率 >= 95%"
+    maxDuration: 600
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-template.yaml

@@ -0,0 +1,72 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve
+# 适用于: 测试 project-evolve skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve"
+description: "Project-Evolve Skill 基准测试 - AI Native 项目自进化"
+
+type: skill
+target: project-evolve
+
+testCases:
+  - name: "init-memory"
+    input:
+      context: "初始化项目记忆"
+    complexity: 1
+    expectedBehaviors:
+      - "扫描项目技术栈"
+      - "检测项目结构"
+      - "自动发现评测点"
+      - "创建记忆目录结构"
+    expectedOutput:
+      - ".autospec/memory/index.yaml"
+      - "初始化报告"
+    successCriteria:
+      - "技术栈识别正确"
+      - "评测点发现 >= 10 个"
+    qualityMetrics:
+      - "记忆创建率 = 100%"
+    maxDuration: 300
+
+  - name: "cruise-mode"
+    input:
+      context: "智能巡航模式"
+    complexity: 3
+    expectedBehaviors:
+      - "加载 validated 记忆"
+      - "执行代码扫描"
+      - "智能优先级排序"
+      - "L1 自动修复"
+    expectedOutput:
+      - "cruise-report.md"
+      - "auto-fixes.json"
+    successCriteria:
+      - "巡航执行完整"
+      - "自动修复正确"
+    qualityMetrics:
+      - "自动修复准确率 >= 90%"
+    maxDuration: 600
+
+  - name: "deep-evolution"
+    input:
+      context: "深度进化模式"
+    complexity: 5
+    expectedBehaviors:
+      - "六层全维度扫描"
+      - "关联分析"
+      - "生成改进方案"
+      - "区分 auto-fixable/manual-review"
+    expectedOutput:
+      - "deep-report.md"
+      - "improvements.md"
+    successCriteria:
+      - "六层扫描完整"
+      - "改进方案可执行"
+    qualityMetrics:
+      - "维度覆盖率 = 100%"
+    maxDuration: 1200
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/project-evolve-verifier-template.yaml

@@ -0,0 +1,51 @@
+# AutoSpec Skill Benchmark Template - Project-Evolve-Verifier
+# 适用于: 测试 project-evolve-verifier skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-project-evolve-verifier"
+description: "Project-Evolve-Verifier Skill 基准测试"
+
+type: skill
+target: project-evolve-verifier
+
+testCases:
+  - name: "fix-verification"
+    input:
+      context: "验证修复效果"
+    complexity: 3
+    expectedBehaviors:
+      - "重新执行失败测试"
+      - "对比修复前后"
+      - "检测退化"
+      - "计算质量分数"
+    expectedOutput:
+      - "验证报告"
+      - "质量分数对比"
+    successCriteria:
+      - "验证执行完整"
+      - "退化检测正确"
+    qualityMetrics:
+      - "验证完整率 = 100%"
+      - "退化检出率 = 100%"
+    maxDuration: 450
+
+  - name: "regression-check"
+    input:
+      context: "回归检测"
+    complexity: 3
+    expectedBehaviors:
+      - "执行全量测试"
+      - "对比历史结果"
+      - "识别退化"
+    expectedOutput:
+      - "回归报告"
+    successCriteria:
+      - "回归检测完整"
+    qualityMetrics:
+      - "退化检出率 = 100%"
+    maxDuration: 600
+
+successCriteria:
+  passRate: 90
+  avgFieldCompletion: 95