@chongyan/autospec 1.0.2 → 1.0.3

package/plugins/benchmarks/templates/skills/benchmark-generator-template.yaml

@@ -1,52 +0,0 @@
-# AutoSpec Skill Benchmark Template - Benchmark-Generator
-# 适用于: 测试 benchmark-generator skill
-# init 后复制到 .autospec/benchmarks/ 后按需修改
-
-version: "1.0"
-name: "skill-benchmark-generator"
-description: "Benchmark-Generator Skill 基准测试"
-
-type: skill
-target: benchmark-generator
-
-testCases:
-  - name: "generate-from-requirement"
-    input:
-      context: "根据需求文档生成 benchmark"
-    complexity: 3
-    expectedBehaviors:
-      - "读取需求文档"
-      - "提取关键场景"
-      - "生成测试用例"
-      - "生成 benchmark YAML"
-    expectedOutput:
-      - "generated-benchmark.yaml"
-      - "测试用例清单"
-    successCriteria:
-      - "场景提取完整"
-      - "benchmark 格式正确"
-    qualityMetrics:
-      - "场景覆盖率 >= 90%"
-      - "格式正确率 = 100%"
-    maxDuration: 300
-
-  - name: "generate-from-design"
-    input:
-      context: "根据设计文档生成 benchmark"
-    complexity: 3
-    expectedBehaviors:
-      - "读取设计文档"
-      - "提取 API 接口"
-      - "生成 API 测试 benchmark"
-    expectedOutput:
-      - "api-benchmark.yaml"
-    successCriteria:
-      - "API 提取完整"
-      - "benchmark 可执行"
-    qualityMetrics:
-      - "API 覆盖率 >= 90%"
-    maxDuration: 300
-
-successCriteria:
-  passRate: 85
-  avgFieldCompletion: 90

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

@@ -1,130 +0,0 @@
-# AutoSpec Skill Benchmark Template - Delivery-Stage
-# 适用于: 测试 delivery-stage skill
-# init 后复制到 .autospec/benchmarks/ 后按需修改
-
-version: "1.0"
-name: "skill-delivery-stage"
-description: "Delivery-Stage Skill 基准测试"
-
-type: skill
-target: delivery-stage
-
-testCases:
-  - name: "simple-delivery"
-    input:
-      context: "交付 {feature-name} 功能"
-    complexity: 1
-    expectedBehaviors:
-      - "执行发布检查清单"
-      - "执行一致性检查"
-      - "生成交付文档"
-      - "执行 Layer1 最终验证"
-    expectedOutput:
-      - "release-notes.md"
-      - "deployment-guide.md"
-      - "consistency-report.md"
-      - "delivery-checklist.md"
-    successCriteria:
-      - "发布检查清单完整执行"
-      - "一致性检查通过"
-      - "交付文档完整"
-    qualityMetrics:
-      - "文档完整率 >= 90%"
-      - "交付清单执行率 = 100%"
-    maxDuration: 300
-
-  - name: "multi-system-delivery"
-    input:
-      context: "交付 {feature} 系统（backend/frontend/mobile）"
-    complexity: 5
-    expectedBehaviors:
-      - "识别所有子系统"
-      - "执行各子系统发布检查"
-      - "执行集成一致性检查"
-      - "生成多系统部署方案"
-      - "生成回滚方案"
-    expectedOutput:
-      - "release-notes.md"
-      - "deployment-plan.md"
-      - "rollback-plan.md"
-      - "integration-report.md"
-    successCriteria:
-      - "所有子系统交付完整"
-      - "部署方案可行"
-      - "回滚方案完整"
-    qualityMetrics:
-      - "子系统覆盖率 = 100%"
-      - "回滚方案完整率 >= 90%"
-    maxDuration: 600
-
-  - name: "ai-delivery"
-    input:
-      context: "交付 AI 功能（含模型）"
-    complexity: 5
-    expectedBehaviors:
-      - "执行模型效果验收"
-      - "执行模型性能验收"
-      - "生成模型卡片"
-      - "生成 A/B 测试方案"
-      - "生成监控方案"
-    expectedOutput:
-      - "model-card.md"
-      - "acceptance-report.md"
-      - "ab-test-plan.md"
-      - "monitoring-plan.md"
-    successCriteria:
-      - "效果指标达标"
-      - "模型卡片完整"
-      - "监控方案完整"
-    qualityMetrics:
-      - "效果验收执行率 = 100%"
-      - "模型卡片完整率 >= 90%"
-    maxDuration: 600
-
-  - name: "safety-audit"
-    input:
-      context: "执行生产数据库迁移"
-    complexity: 3
-    expectedBehaviors:
-      - "执行安全审计"
-      - "验证数据备份方案"
-      - "验证回滚方案"
-      - "生成人工确认清单"
-    expectedOutput:
-      - "safety-audit.md"
-      - "backup-plan.md"
-      - "rollback-plan.md"
-      - "manual-checklist.md"
-    successCriteria:
-      - "安全审计执行"
-      - "备份方案验证通过"
-      - "人工确认清单完整"
-    qualityMetrics:
-      - "安全审计执行率 = 100%"
-      - "人工确认清单完整率 = 100%"
-    maxDuration: 450
-
-  - name: "practice-log"
-    input:
-      context: "记录交付后的实践日志和进化信号"
-    complexity: 3
-    expectedBehaviors:
-      - "收集实践日志"
-      - "检测进化信号"
-      - "生成进化建议"
-      - "更新 practice-log.json"
-    expectedOutput:
-      - "practice-log.json"
-      - "evolution-signal.md"
-      - "improvement-suggestion.md"
-    successCriteria:
-      - "实践日志格式正确"
-      - "进化信号检测正确"
-    qualityMetrics:
-      - "实践日志完整率 >= 90%"
-      - "进化信号检出率 = 100%"
-    maxDuration: 300
-
-successCriteria:
-  passRate: 85
-  avgFieldCompletion: 90

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

@@ -1,131 +0,0 @@
-# AutoSpec Skill Benchmark Template - Design-Stage
-# 适用于: 测试 design-stage skill
-# init 后复制到 .autospec/benchmarks/ 后按需修改
-
-version: "1.0"
-name: "skill-design-stage"
-description: "Design-Stage Skill 基准测试"
-
-type: skill
-target: design-stage
-
-testCases:
-  - name: "simple-design"
-    input:
-      context: "{feature-name} 功能设计"
-    complexity: 1
-    expectedBehaviors:
-      - "执行架构设计（模块划分、接口定义）"
-      - "设计 API 接口（RESTful 规范）"
-      - "设计数据库模型（ER 图、表结构）"
-      - "定义安全机制（认证、授权、加密）"
-    expectedOutput:
-      - "design.md"
-      - "api.yaml"
-      - "db-schema.sql"
-      - "security-design.md"
-    successCriteria:
-      - "架构设计合理"
-      - "API 设计符合 RESTful 规范"
-      - "数据库设计满足范式要求"
-    qualityMetrics:
-      - "设计完整率 >= 90%"
-      - "API 规范符合率 >= 95%"
-    maxDuration: 600
-
-  - name: "multi-system-design"
-    input:
-      context: "{feature} 设计，包含 {subsystems}"
-    complexity: 5
-    expectedBehaviors:
-      - "识别所有子系统"
-      - "设计整体架构"
-      - "定义系统间契约"
-      - "设计依赖顺序"
-      - "执行团队对抗审查"
-    expectedOutput:
-      - "design.md"
-      - "backend.md"
-      - "frontend.md"
-      - "contracts/api.yaml"
-    successCriteria:
-      - "识别所有子系统"
-      - "系统间契约定义清晰"
-      - "依赖顺序正确"
-    qualityMetrics:
-      - "子系统识别率 = 100%"
-      - "契约完整率 >= 90%"
-    maxDuration: 1200
-
-  - name: "ai-design"
-    input:
-      context: "AI 功能设计，使用 {model-type}"
-    complexity: 5
-    expectedBehaviors:
-      - "识别 AI/模型组件"
-      - "设计模型选型方案"
-      - "设计效果评测方案"
-      - "设计推理服务架构"
-      - "设计降级方案"
-    expectedOutput:
-      - "design.md"
-      - "ai-model.md"
-      - "evaluation.md"
-      - "fallback.md"
-    successCriteria:
-      - "AI 组件设计清晰"
-      - "效果指标可测量"
-      - "降级方案完整"
-    qualityMetrics:
-      - "AI 组件完整率 >= 90%"
-      - "评测方案完整率 >= 90%"
-    maxDuration: 900
-
-  - name: "data-modeling-design"
-    input:
-      context: "数据仓库设计，支持 BI 分析"
-    complexity: 5
-    expectedBehaviors:
-      - "识别数据源"
-      - "设计数据模型（维度建模）"
-      - "设计 ETL 流程"
-      - "设计数据质量检查机制"
-    expectedOutput:
-      - "design.md"
-      - "data-model.md"
-      - "etl-pipeline.md"
-      - "data-quality.md"
-    successCriteria:
-      - "数据源识别完整"
-      - "ETL 流程清晰可执行"
-    qualityMetrics:
-      - "数据源完整率 >= 90%"
-      - "ETL 完整率 >= 90%"
-    maxDuration: 900
-
-  - name: "high-concurrency-design"
-    input:
-      context: "秒杀系统设计，支持 10 万 QPS"
-    complexity: 5
-    expectedBehaviors:
-      - "设计缓存策略"
-      - "设计限流降级方案"
-      - "设计异步处理"
-      - "设计数据库优化"
-      - "设计容量规划"
-    expectedOutput:
-      - "design.md"
-      - "cache-design.md"
-      - "rate-limiting.md"
-      - "scaling.md"
-    successCriteria:
-      - "缓存策略合理"
-      - "限流降级方案完整"
-    qualityMetrics:
-      - "性能考虑完整率 >= 90%"
-      - "高可用设计合理率 >= 90%"
-    maxDuration: 900
-
-successCriteria:
-  passRate: 85
-  avgFieldCompletion: 90

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

@@ -1,60 +0,0 @@
-# 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

@@ -1,114 +0,0 @@
-# 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

@@ -1,51 +0,0 @@
-# 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

@@ -1,34 +0,0 @@
-# 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

@@ -1,50 +0,0 @@
-# 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

@@ -1,52 +0,0 @@
-# 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