@chongyan/autospec 1.0.1 → 1.0.2

package/plugins/benchmarks/templates/skills/experiment-iteration-template.yaml

@@ -0,0 +1,60 @@
+# AutoSpec Skill Benchmark Template - Experiment-Iteration
+# 适用于: 测试 experiment-iteration skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-experiment-iteration"
+description: "Experiment-Iteration Skill 基准测试 - 实验模式验证流程"
+
+type: skill
+target: experiment-iteration
+
+testCases:
+  - name: "simple-experiment"
+    input:
+      context: "验证 {technology} 用于 {use-case} 的可行性"
+    complexity: 3
+    expectedBehaviors:
+      - "定义假设"
+      - "设计实验"
+      - "实现原型"
+      - "执行评测"
+      - "验证假设"
+    expectedOutput:
+      - "hypothesis.md"
+      - "prototype 代码"
+      - "evaluation-report.md"
+      - "conclusion.md"
+    successCriteria:
+      - "假设定义清晰"
+      - "评测执行完整"
+      - "结论明确"
+    qualityMetrics:
+      - "假设验证完整率 = 100%"
+      - "评测执行率 = 100%"
+    maxDuration: 1800
+
+  - name: "ai-experiment"
+    input:
+      context: "验证 AI 模型效果"
+    complexity: 5
+    expectedBehaviors:
+      - "定义效果指标"
+      - "构建评测数据集"
+      - "训练/微调模型"
+      - "执行效果评测"
+    expectedOutput:
+      - "evaluation-plan.md"
+      - "dataset.jsonl"
+      - "evaluation-report.md"
+    successCriteria:
+      - "效果指标可测量"
+      - "评测数据集有代表性"
+    qualityMetrics:
+      - "指标完整率 >= 90%"
+      - "数据集覆盖率 >= 80%"
+    maxDuration: 3600
+
+successCriteria:
+  passRate: 80
+  avgFieldCompletion: 85

package/plugins/benchmarks/templates/skills/exploration-phase-template.yaml

@@ -0,0 +1,114 @@
+# AutoSpec Skill Benchmark Template - Exploration-Phase
+# 适用于: 测试 exploration-phase skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-exploration-phase"
+description: "Exploration-Phase Skill 基准测试"
+
+type: skill
+target: exploration-phase
+
+testCases:
+  - name: "simple-exploration"
+    input:
+      context: "实现 {feature-name} 功能"
+    complexity: 1
+    expectedBehaviors:
+      - "执行复杂度评估（simple）"
+      - "启动 2 个 Agent（CEO + 产品负责人）"
+      - "执行 2-3 轮澄清"
+      - "每轮最多 5 个问题"
+    expectedOutput:
+      - "clarifications.md"
+      - "requirement.md"
+      - "复杂度评估结果"
+    successCriteria:
+      - "澄清轮次 2-3 轮"
+      - "问题数 <= 5 个/轮"
+      - "需求文档结构完整"
+    qualityMetrics:
+      - "字段完整率 >= 90%"
+      - "澄清问题质量 >= 80%"
+    maxDuration: 600
+
+  - name: "medium-exploration"
+    input:
+      context: "实现 {feature} 功能，包含多个模块"
+    complexity: 3
+    expectedBehaviors:
+      - "执行复杂度评估（medium）"
+      - "启动 3 个 Agent（CEO + 产品 + 技术）"
+      - "执行 3-4 轮澄清"
+      - "识别技术风险"
+    expectedOutput:
+      - "clarifications.md"
+      - "requirement.md"
+      - "技术风险评估"
+    successCriteria:
+      - "澄清轮次 3-4 轮"
+      - "技术风险识别完整"
+    qualityMetrics:
+      - "风险识别率 >= 90%"
+    maxDuration: 900
+
+  - name: "complex-exploration"
+    input:
+      context: "实现 {feature}，包含 AI 模型、后端 API、前端界面"
+    complexity: 5
+    expectedBehaviors:
+      - "执行复杂度评估（complex）"
+      - "启动 4+ 个 Agent"
+      - "执行 5-6 轮澄清"
+      - "识别多系统边界"
+      - "定义系统间契约"
+    expectedOutput:
+      - "clarifications.md"
+      - "requirement.md"
+      - "多系统需求分析"
+      - "系统间契约定义"
+    successCriteria:
+      - "多系统识别完整"
+      - "系统间契约清晰"
+    qualityMetrics:
+      - "系统识别率 = 100%"
+      - "契约完整率 >= 90%"
+    maxDuration: 1200
+
+  - name: "ambiguous-requirement"
+    input:
+      context: "做一个好用的后台管理系统"
+    complexity: 3
+    expectedBehaviors:
+      - "识别需求歧义"
+      - "执行结构化澄清"
+      - "每个问题附 AI 推荐答案"
+      - "支持用户提前终止"
+    successCriteria:
+      - "歧义识别完整"
+      - "澄清问题 <= 5 个"
+      - "每个问题有推荐答案"
+    qualityMetrics:
+      - "歧义识别率 >= 90%"
+      - "推荐答案合理率 >= 90%"
+    maxDuration: 600
+
+  - name: "early-termination"
+    input:
+      context: "实现 {feature} 功能"
+      earlyTermination: true
+    complexity: 1
+    expectedBehaviors:
+      - "执行第 1 轮澄清"
+      - "正确识别用户终止意图"
+      - "提前终止澄清流程"
+    successCriteria:
+      - "终止识别准确"
+      - "需求文档完整"
+    qualityMetrics:
+      - "终止识别准确率 = 100%"
+    maxDuration: 300
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

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

@@ -0,0 +1,51 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Analyzer
+# 适用于: 测试 field-evolve-analyzer skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-analyzer"
+description: "Field-Evolve-Analyzer Skill 基准测试"
+
+type: skill
+target: field-evolve-analyzer
+
+testCases:
+  - name: "problem-analysis"
+    input:
+      context: "分析测试失败的问题"
+    complexity: 3
+    expectedBehaviors:
+      - "读取测试结果"
+      - "识别失败模式"
+      - "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:
+      - "失败模式识别准确"
+      - "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/field-evolve-distiller-template.yaml

@@ -0,0 +1,34 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Distiller
+# 适用于: 测试 field-evolve-distiller skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-distiller"
+description: "Field-Evolve-Distiller Skill 基准测试"
+
+type: skill
+target: field-evolve-distiller
+
+testCases:
+  - name: "skill-distillation"
+    input:
+      context: "从实践日志提炼技能"
+    complexity: 3
+    expectedBehaviors:
+      - "读取进化规则"
+      - "收集 practice-log"
+      - "执行技能蒸馏"
+      - "验证进化效果"
+    expectedOutput:
+      - "distill-report.md"
+      - "distilled-skills/"
+    successCriteria:
+      - "practice-log 完整读取"
+      - "技能提炼符合规范"
+    qualityMetrics:
+      - "技能规范率 >= 90%"
+    maxDuration: 600
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

package/plugins/benchmarks/templates/skills/field-evolve-executor-template.yaml

@@ -0,0 +1,50 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Executor
+# 适用于: 测试 field-evolve-executor skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-executor"
+description: "Field-Evolve-Executor Skill 基准测试"
+
+type: skill
+target: field-evolve-executor
+
+testCases:
+  - name: "benchmark-execution"
+    input:
+      context: "执行 benchmark 测试"
+    complexity: 3
+    expectedBehaviors:
+      - "读取 benchmark YAML"
+      - "创建临时目录"
+      - "执行完整流程"
+      - "捕获详细指标"
+    expectedOutput:
+      - "执行结果"
+      - "指标数据"
+      - "产出物"
+    successCriteria:
+      - "benchmark 执行完整"
+      - "指标捕获准确"
+    qualityMetrics:
+      - "benchmark 执行率 = 100%"
+    maxDuration: 900
+
+  - name: "isolated-execution"
+    input:
+      context: "在隔离环境中执行测试"
+    complexity: 5
+    expectedBehaviors:
+      - "创建临时 git worktree"
+      - "执行测试"
+      - "清理环境"
+    successCriteria:
+      - "隔离环境创建成功"
+      - "清理完整"
+    qualityMetrics:
+      - "隔离成功率 = 100%"
+    maxDuration: 600
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

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

@@ -0,0 +1,52 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Fixer
+# 适用于: 测试 field-evolve-fixer skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-fixer"
+description: "Field-Evolve-Fixer Skill 基准测试"
+
+type: skill
+target: field-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/field-evolve-learner-template.yaml

@@ -0,0 +1,33 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Learner
+# 适用于: 测试 field-evolve-learner skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-learner"
+description: "Field-Evolve-Learner Skill 基准测试"
+
+type: skill
+target: field-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
+
+successCriteria:
+  passRate: 80
+  avgFieldCompletion: 85

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

@@ -0,0 +1,74 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Scanner
+# 适用于: 测试 field-evolve-scanner skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-scanner"
+description: "Field-Evolve-Scanner Skill 基准测试"
+
+type: skill
+target: field-evolve-scanner
+
+testCases:
+  - name: "simple-project-scan"
+    input:
+      context: "扫描一个简单的 Node.js 项目"
+    complexity: 1
+    expectedBehaviors:
+      - "检测技术栈"
+      - "检测项目结构"
+      - "识别代码复杂度问题"
+      - "输出问题清单"
+    expectedOutput:
+      - "项目特征报告"
+      - "问题清单"
+    successCriteria:
+      - "技术栈检测正确"
+      - "问题识别完整"
+    qualityMetrics:
+      - "技术栈识别准确率 >= 95%"
+      - "问题识别率 >= 80%"
+    maxDuration: 300
+
+  - name: "multi-system-scan"
+    input:
+      context: "扫描包含前后端的多系统项目"
+    complexity: 5
+    expectedBehaviors:
+      - "检测多个子系统"
+      - "识别系统边界"
+      - "分析依赖关系"
+      - "启用兜底扫描分类文档"
+    expectedOutput:
+      - "多系统特征报告"
+      - "文档分类报告"
+    successCriteria:
+      - "子系统识别完整"
+      - "文档分类准确"
+    qualityMetrics:
+      - "子系统识别率 = 100%"
+      - "文档分类准确率 >= 90%"
+    maxDuration: 600
+
+  - name: "fallback-classification"
+    input:
+      context: "扫描项目并分类未匹配规则的文档"
+    complexity: 3
+    expectedBehaviors:
+      - "收集所有 Markdown 文件"
+      - "排除已规则匹配的文件"
+      - "使用 AI 模型分类"
+    expectedOutput:
+      - "设计文档列表"
+      - "业务文档列表"
+      - "测试文档列表"
+    successCriteria:
+      - "分类逻辑正确"
+      - "分类结果合理"
+    qualityMetrics:
+      - "分类准确率 >= 85%"
+    maxDuration: 450
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

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

@@ -0,0 +1,71 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve
+# 适用于: 测试 field-evolve skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve"
+description: "Field-Evolve Skill 基准测试 - 实战项目自进化"
+
+type: skill
+target: field-evolve
+
+testCases:
+  - name: "quick-validation"
+    input:
+      context: "快速验证模式"
+    complexity: 1
+    expectedBehaviors:
+      - "读取 config.json"
+      - "执行编译检查"
+      - "执行单元测试"
+      - "执行 Lint 检查"
+    expectedOutput:
+      - "test-result.json"
+      - "quick-report.md"
+    successCriteria:
+      - "所有检查项执行"
+      - "结果记录完整"
+    qualityMetrics:
+      - "检查执行率 = 100%"
+    maxDuration: 300
+
+  - name: "deep-testing"
+    input:
+      context: "深度测试模式"
+    complexity: 5
+    expectedBehaviors:
+      - "扫描 benchmarks"
+      - "执行 benchmark 测试场景"
+      - "计算三维度评分"
+    expectedOutput:
+      - "deep-report.md"
+      - "evaluation.json"
+    successCriteria:
+      - "benchmark 执行完整"
+      - "评分计算正确"
+    qualityMetrics:
+      - "benchmark 执行率 = 100%"
+    maxDuration: 1800
+
+  - name: "full-cycle"
+    input:
+      context: "完整循环模式"
+    complexity: 5
+    expectedBehaviors:
+      - "执行深度测试"
+      - "生成改进方案"
+      - "执行自动修复"
+      - "技能蒸馏"
+    expectedOutput:
+      - "full-report.md"
+      - "distilled-skills/"
+    successCriteria:
+      - "完整循环执行成功"
+      - "无退化发生"
+    qualityMetrics:
+      - "修复有效率 >= 85%"
+    maxDuration: 2400
+
+successCriteria:
+  passRate: 85
+  avgFieldCompletion: 90

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

@@ -0,0 +1,51 @@
+# AutoSpec Skill Benchmark Template - Field-Evolve-Verifier
+# 适用于: 测试 field-evolve-verifier skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-field-evolve-verifier"
+description: "Field-Evolve-Verifier Skill 基准测试"
+
+type: skill
+target: field-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: "benchmark-comparison"
+    input:
+      context: "对比 benchmark 结果"
+    complexity: 3
+    expectedBehaviors:
+      - "执行 benchmark"
+      - "对比历史结果"
+      - "识别退化"
+    expectedOutput:
+      - "对比报告"
+    successCriteria:
+      - "对比准确"
+    qualityMetrics:
+      - "对比准确率 = 100%"
+    maxDuration: 600
+
+successCriteria:
+  passRate: 90
+  avgFieldCompletion: 95

package/plugins/benchmarks/templates/skills/hotfix-iteration-template.yaml

@@ -0,0 +1,54 @@
+# AutoSpec Skill Benchmark Template - Hotfix-Iteration
+# 适用于: 测试 hotfix-iteration skill
+# init 后复制到 .autospec/benchmarks/ 后按需修改
+
+version: "1.0"
+name: "skill-hotfix-iteration"
+description: "Hotfix-Iteration Skill 基准测试 - 热修复快速流程"
+
+type: skill
+target: hotfix-iteration
+
+testCases:
+  - name: "simple-hotfix"
+    input:
+      context: "修复生产环境 {bug-description}"
+    complexity: 1
+    expectedBehaviors:
+      - "问题诊断"
+      - "风险评估"
+      - "快速修复"
+      - "快速审查"
+      - "快速部署"
+    expectedOutput:
+      - "hotfix-issue.md"
+      - "变更代码"
+      - "回归测试"
+    successCriteria:
+      - "问题定位准确"
+      - "变更最小化"
+      - "回归测试通过"
+    qualityMetrics:
+      - "修复耗时 < 15 分钟"
+      - "变更行数 < 50"
+    maxDuration: 900
+
+  - name: "critical-hotfix"
+    input:
+      context: "修复生产环境严重安全漏洞"
+    complexity: 3
+    expectedBehaviors:
+      - "安全评估"
+      - "影响范围分析"
+      - "紧急修复"
+      - "安全审查"
+    successCriteria:
+      - "安全风险识别完整"
+      - "修复方案合理"
+    qualityMetrics:
+      - "修复耗时 < 30 分钟"
+    maxDuration: 1800
+
+successCriteria:
+  passRate: 95
+  avgFieldCompletion: 95