@chongyan/autospec 1.0.1

package/plugins/memory/templates/team-habits.yaml

@@ -0,0 +1,119 @@
+# AutoSpec Memory Template - Team Habits
+# 团队习惯学习记忆模板
+
+version: "1.0"
+memory:
+  id: "template-team-habits"
+  name: "团队习惯学习模板"
+  description: "学习团队编码习惯和偏好"
+  category: "team-habits"
+
+  # 此模板主要用于学习，而非硬性规则
+  type: "learning"
+
+  # 习惯维度
+  habit_dimensions:
+    # 维度1: 代码风格
+    - id: "code-style"
+      name: "代码风格偏好"
+      description: "团队代码风格习惯"
+
+      observations:
+        - aspect: "indentation"
+          options: ["2-spaces", "4-spaces", "tab"]
+          detection: "analyze-existing-code"
+
+        - aspect: "quotes"
+          options: ["single", "double"]
+          detection: "analyze-existing-code"
+
+        - aspect: "semicolons"
+          options: ["always", "never"]
+          detection: "analyze-existing-code"
+
+        - aspect: "trailing-commas"
+          options: ["always", "never", "es5"]
+          detection: "analyze-existing-code"
+
+    # 维度2: 命名偏好
+    - id: "naming"
+      name: "命名偏好"
+      description: "团队命名习惯"
+
+      observations:
+        - aspect: "boolean-prefix"
+          options: ["is-", "has-", "can-", "should-", "no-prefix"]
+          examples: ["isActive", "hasPermission"]
+
+        - aspect: "collection-suffix"
+          options: ["-list", "-array", "-collection", "plural", "no-suffix"]
+          examples: ["userList", "users"]
+
+        - aspect: "abbreviation"
+          options: ["full-words", "common-abbrev", "any-abbrev"]
+          examples: ["configuration" vs "config"]
+
+    # 维度3: 架构偏好
+    - id: "architecture"
+      name: "架构偏好"
+      description: "团队架构设计习惯"
+
+      observations:
+        - aspect: "function-style"
+          options: ["arrow-functions", "function-declaration", "mixed"]
+
+        - aspect: "class-usage"
+          options: ["classes", "functions-only", "mixed"]
+
+        - aspect: "type-strictness"
+          options: ["strict", "moderate", "loose"]
+
+    # 维度4: 自动化接受度
+    - id: "automation"
+      name: "自动化接受度"
+      description: "团队对自动化的接受程度"
+
+      observations:
+        - aspect: "auto-fix-acceptance"
+          detection: "track-suggestion-acceptance-rate"
+
+        - aspect: "preferred-notification"
+          options: ["immediate", "daily-summary", "weekly-summary", "on-demand"]
+
+        - aspect: "confirmation-threshold"
+          description: "什么级别的修改需要确认"
+          options: ["all", "non-trivial", "high-impact-only", "none"]
+
+  # 学习机制
+  learning:
+    # 观察周期
+    observation_period_days: 14
+
+    # 学习来源
+    sources:
+      - "existing-codebase"
+      - "manual-changes"
+      - "feedback-explicit"
+      - "feedback-implicit"
+
+    # 置信度阈值
+    confidence_threshold: 0.8
+
+    # 应用方式
+    application:
+      - "adjust-rule-weights"
+      - "prioritize-suggestions"
+      - "customize-auto-fixes"
+      - "filter-noise"
+
+  # 反馈收集
+  feedback:
+    explicit:
+      - "rating-suggestions"
+      - "accept-reject-tracking"
+      - "custom-rule-requests"
+
+    implicit:
+      - "suggestion-acceptance-rate"
+      - "time-to-accept"
+      - "manual-override-patterns"

package/plugins/memory/templates/testing.yaml

@@ -0,0 +1,121 @@
+# AutoSpec Memory Template - Testing
+# 测试体系评测记忆模板
+
+version: "1.0"
+memory:
+  id: "template-testing"
+  name: "测试体系评测模板"
+  description: "测试覆盖率、质量、稳定性评测"
+  category: "testing"
+
+  rules:
+    # 规则1: 测试覆盖率缺口
+    - id: "coverage-gap"
+      name: "测试覆盖率缺口"
+      description: "识别变更频繁但测试覆盖低的代码"
+      severity: "warning"
+
+      detection:
+        type: "coverage-analysis"
+        thresholds:
+          line_coverage: 80
+          branch_coverage: 70
+          function_coverage: 80
+
+        # 重点关注：变更频繁 + 覆盖低
+        hot_spot_analysis:
+          min_change_frequency: 3  # 过去30天变更次数
+          max_coverage: 60
+
+      fixes:
+        - type: "suggest"
+          description: "生成缺失测试模板"
+          action: "generate-test-template"
+
+    # 规则2: 边界测试
+    - id: "boundary-testing"
+      name: "边界条件测试"
+      description: "检查是否缺少边界条件测试"
+      severity: "warning"
+
+      detection:
+        type: "boundary-analysis"
+        patterns:
+          - "array-index-without-bound-check"
+          - "division-without-zero-check"
+          - "null-check-missing"
+          - "empty-string-not-tested"
+
+      fixes:
+        - type: "suggest"
+          description: "建议边界测试用例"
+          action: "suggest-boundary-tests"
+
+    # 规则3: Flaky Test检测
+    - id: "flaky-test"
+      name: "不稳定测试检测"
+      description: "检测执行不稳定的测试"
+      severity: "error"
+
+      detection:
+        type: "test-history"
+        criteria:
+          min_runs: 10
+          max_failure_rate: 0.1
+          max_duration_variance: 0.5
+
+      fixes:
+        - type: "suggest"
+          description: "分析不稳定原因并建议修复"
+          action: "analyze-flaky-cause"
+
+    # 规则4: Mock使用
+    - id: "mock-usage"
+      name: "Mock使用检查"
+      description: "检查Mock使用是否恰当"
+      severity: "info"
+
+      detection:
+        type: "mock-analysis"
+        checks:
+          - "mocking-internal-functions"  # 不应该Mock内部函数
+          - "incomplete-mock"              # Mock不完整
+          - "over-mocking"                 # 过度Mock
+
+      fixes:
+        - type: "suggest"
+          description: "建议Mock改进方案"
+          action: "suggest-mock-improvement"
+
+    # 规则5: 测试数据
+    - id: "test-data"
+      name: "测试数据管理"
+      description: "检查测试数据管理是否规范"
+      severity: "info"
+
+      detection:
+        type: "test-data-analysis"
+        checks:
+          - "hardcoded-test-data"
+          - "missing-test-fixtures"
+          - "inconsistent-test-data"
+
+      fixes:
+        - type: "auto"
+          description: "提取测试数据到fixtures"
+          action: "extract-test-fixtures"
+
+  execution:
+    order:
+      - "flaky-test"       # 优先处理不稳定测试
+      - "coverage-gap"
+      - "boundary-testing"
+      - "mock-usage"
+      - "test-data"
+
+    conditions:
+      - "test-pass"
+
+    on_failure:
+      action: "report"
+      notify: true

package/plugins/skills/README.en.md

@@ -0,0 +1,47 @@
+# Skills Template Directory
+
+This directory contains template definitions for all Skills, which are copied to `.claude/skills/` during `autospec init`.
+
+## Directory Structure
+
+```
+skills/
+├── startup-guard/
+│   └── SKILL.md                   # Startup gate (auto-triggered)
+├── layer1-validation/
+│   └── SKILL.md                   # Layer 1 validation (auto-triggered)
+├── requirement-stage/
+│   └── SKILL.md                   # Requirement analysis stage
+├── design-stage/
+│   └── SKILL.md                   # Design planning stage
+├── implementation-stage/
+│   └── SKILL.md                   # Code implementation stage
+├── testing-stage/
+│   └── SKILL.md                   # QA testing stage
+├── delivery-stage/
+│   └── SKILL.md                   # Delivery and deployment stage
+├── evolution-process/
+│   └── SKILL.md                   # Self-evolution (auto-triggered)
+├── pending-dashboard/
+│   └── SKILL.md                   # Pending dashboard (session startup trigger)
+└── README.md                      # This file
+```
+
+## Skill Categories
+
+**Auto-triggered**:
+- `startup-guard`: Auto-triggered when development task detected
+- `layer1-validation`: Auto-triggered after code changes
+- `evolution-process`: Auto-triggered after pipeline completion
+- `pending-dashboard`: Auto-triggered on session startup
+
+**Stage-based**:
+- `requirement-stage`: Requirement analysis stage
+- `design-stage`: Design planning stage
+- `implementation-stage`: Code implementation stage
+- `testing-stage`: QA testing stage
+- `delivery-stage`: Delivery and deployment stage
+
+## Variable Substitution
+
+The `${KNOWLEDGE}` variable in templates is replaced with `.autospec/knowledge` during copying.

package/plugins/skills/README.md

@@ -0,0 +1,104 @@
+# Skills 模板目录
+
+本目录包含所有 Skill 技能的模板定义，在 `autospec init` 时复制到 `.claude/skills/` 目录。
+
+## 人类-[skills]-创业团队-skills-领域知识分层
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│  人类维护层 (Human-maintained)                                    │
+│  ../knowledge/domain/ - 业务知识（术语、规则、流程）                │
+│  由领域专家维护，AI 读取但不修改                                   │
+└─────────────────────────────────────────────────────────────────┘
+                              ↓ 驱动
+┌─────────────────────────────────────────────────────────────────┐
+│  创业团队 Skills (AI Agent Roles)                                 │
+│  ../agents/roles/ - CEO、产品负责人、技术负责人、工程师等           │
+│  模拟创业团队对抗审查，多视角审视产出                              │
+└─────────────────────────────────────────────────────────────────┘
+                              ↓ 执行
+┌─────────────────────────────────────────────────────────────────┐
+│  阶段 Skills (Stage Skills)                                       │
+│  本目录 *-stage/ - 需求分析、方案设计、编码实现等阶段技能           │
+│  驱动各阶段执行，产出标准化产物                                    │
+└─────────────────────────────────────────────────────────────────┘
+                              ↓ 支撑
+┌─────────────────────────────────────────────────────────────────┐
+│  领域知识/规范分层 (Knowledge & Standards)                        │
+│  ../knowledge/ - 原则、流程、指南、检查清单                         │
+│  按冻结区/受控区/自由区分层管理                                    │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+## 目录结构
+
+```
+skills/
+├── startup-guard/
+│   └── SKILL.md                   # 启动门禁（自动触发）
+├── layer1-validation/
+│   └── SKILL.md                   # Layer 1 验证（自动触发）
+├── requirement-stage/
+│   └── SKILL.md                   # 需求分析阶段（含 Benchmarks 评测）
+├── design-stage/
+│   └── SKILL.md                   # 方案设计阶段（含 Benchmarks 评测）
+├── implementation-stage/
+│   └── SKILL.md                   # 编码实现阶段（含 Benchmarks 评测）
+├── testing-stage/
+│   └── SKILL.md                   # QA测试阶段（含 Benchmarks 评测）
+├── delivery-stage/
+│   └── SKILL.md                   # 交付部署阶段
+├── evolution-process/
+│   └── SKILL.md                   # 自进化（自动触发）- L2 Field-Evolve
+├── project-evolution/
+│   └── SKILL.md                   # 项目进化 - L3 Project-Evolve
+├── benchmark-generator/
+│   ├── SKILL.md                   # Benchmarks 自动生成
+│   └── README.md                  # 使用文档
+├── benchmark-executor/
+│   ├── SKILL.md                   # Benchmarks 执行评测
+│   └── README.md                  # 使用文档
+├── skill-forge/
+│   ├── SKILL.md                   # 技能锻造系统（蒸馏/迭代/优化）
+│   └── references/                # 参考文档
+├── pending-dashboard/
+│   └── SKILL.md                   # 待办仪表盘（会话启动触发）
+└── README.md                      # 本文件
+```
+
+## Skill 分类
+
+### 按触发方式分类
+
+**自动触发型**：
+- `startup-guard`：检测到开发任务时自动触发
+- `layer1-validation`：代码变更后自动触发
+- `evolution-process`：流程完成后自动触发（L2 Field-Evolve）
+- `pending-dashboard`：会话启动时自动触发
+
+**阶段流程型**：
+- `requirement-stage`：需求分析阶段（含 Layer 3 Benchmarks 评测）
+- `design-stage`：方案设计阶段（含 Layer 3 Benchmarks 评测）
+- `implementation-stage`：编码实现阶段（含 Layer 3 Benchmarks 评测）
+- `testing-stage`：QA测试阶段（含 Layer 3 Benchmarks 评测）
+- `delivery-stage`：交付部署阶段
+
+**评测驱动型**：
+- `benchmark-generator`：根据项目特点自动生成阶段 benchmarks
+- `benchmark-executor`：执行 benchmarks 评测，生成改进建议
+
+**进化支撑型**：
+- `skill-forge`：技能锻造系统（蒸馏新建/迭代进化/Description优化）
+- `project-evolution`：项目自我进化（L3 Project-Evolve）
+
+### 按三层自进化体系分类
+
+| 层级 | 命令 | Skill |
+|------|------|-------|
+| L1: Self-Evolve | `/self-evolve` | `self-evolve/*`（位于 `.claude/skills/`） |
+| L2: Field-Evolve | `/autospec:field-evolve` | `evolution-process` |
+| L3: Project-Evolve | `/autospec:project-evolve` | `project-evolution` |
+
+## 变量替换
+
+模板中的 `${KNOWLEDGE}` 变量在复制时会被替换为 `.autospec/knowledge`。

package/plugins/skills/benchmark-executor/README.md

@@ -0,0 +1,93 @@
+# Benchmarks 执行器
+
+实战流程 AI 评测系统的核心执行组件。
+
+## 定位
+
+在 AutoSpec 实战流程各阶段中执行 benchmarks 评测，对比实际产出与预期标准，生成评测报告指导优化迭代。
+
+## 核心组件
+
+### 1. BenchmarkEvaluator
+
+负责加载 benchmarks 配置，执行评测，计算评分，生成报告。
+
+```javascript
+const evaluator = new BenchmarkEvaluator(projectRoot);
+evaluator.init();
+
+// 执行评测
+const report = await evaluator.evaluate('design');
+```
+
+### 2. ChangeDetector
+
+负责检测阶段产出物变更，自动触发重新评测。
+
+```javascript
+const detector = new ChangeDetector(projectRoot);
+detector.init();
+
+// 监听变更
+detector.on('change', (changes) => {
+  console.log('检测到变更:', changes);
+});
+
+detector.watch('design');
+```
+
+## 三层验证体系
+
+| 层级 | 验证类型 | 执行方式 | 用途 |
+|------|---------|---------|------|
+| Layer 1 | 确定性验证 | 自动执行 | Schema校验、规则检查、编译测试 |
+| Layer 2 | AI 审查 | Agent 执行 | 质量评估、一致性检查、深度分析 |
+| Layer 3 | Benchmarks 评测 | 自动+AI | 综合评分、改进建议、准出门禁 |
+
+## 与阶段技能的集成
+
+每个阶段技能在验证部分增加 Layer 3 Benchmarks 评测：
+
+- **requirement-stage**: REQ-001~003 需求完整性评测
+- **design-stage**: DES-001~004 设计规范性评测
+- **implementation-stage**: IMP-001~003 代码质量评测
+- **testing-stage**: TST-001~003 测试充分性评测
+
+## 触发方式
+
+1. **自动触发**: 阶段产出物生成/变更后自动执行
+2. **手动触发**: `/benchmark --run --stage={stage}`
+3. **定时触发**: CI 集成，提交前/合并前自动执行
+
+## 评测结果
+
+```json
+{
+  "run_id": "20240326-001",
+  "stage": "design",
+  "summary": {
+    "total": 4,
+    "passed": 3,
+    "failed": 1,
+    "overall_score": 87,
+    "grade": "B+"
+  },
+  "results": [...],
+  "recommendations": [...]
+}
+```
+
+## 准出标准
+
+- 所有 P0 benchmark 必须通过
+- 综合得分 >= 80
+- 无 critical 级别问题
+
+## 改进建议
+
+评测未通过时，自动生成：
+- 问题描述
+- 改进建议
+- 参考示例
+- 预计工作量
+- 是否可自动修复