ssd-ql-workflow 0.2.2 → 0.4.0

package/bin/cli.js

@@ -21,9 +21,11 @@ program
   .option('-f, --force', 'Overwrite existing files', false)
   .option('--skip-commands', 'Skip copying command files', false)
   .option('--skip-schema', 'Skip copying schema files', false)
+  .option('--schema <name>', 'Schema to use: trinity-workflow | hybrid-workflow', 'hybrid-workflow')
   .action(async (options) => {
     const cwd = process.cwd();
     console.log(chalk.blue('\n🚀 Initializing SDD workflow...\n'));
+    console.log(chalk.gray(`Schema: ${options.schema}\n`));
 
     try {
       // 1. Copy openspec config and schema
@@ -31,7 +33,7 @@ program
         const openspecDir = path.join(cwd, 'openspec');
 
         // Create openspec directory structure
-        await fs.ensureDir(path.join(openspecDir, 'schemas', 'trinity-workflow'));
+        await fs.ensureDir(path.join(openspecDir, 'schemas', options.schema));
         await fs.ensureDir(path.join(openspecDir, 'specs'));
         await fs.ensureDir(path.join(openspecDir, 'changes'));
 
@@ -42,28 +44,31 @@ program
         if (await fs.exists(configDest) && !options.force) {
           console.log(chalk.yellow('⚠ config.yaml already exists, use --force to overwrite'));
         } else {
-          await fs.copy(configSrc, configDest);
+          // Update config.yaml with selected schema
+          let configContent = await fs.readFile(configSrc, 'utf8');
+          configContent = configContent.replace(/schema: .*/, `schema: ${options.schema}`);
+          await fs.writeFile(configDest, configContent);
           console.log(chalk.green('✓ Created openspec/config.yaml'));
         }
 
-        // Copy trinity-workflow schema
-        const schemaSrc = path.join(TEMPLATES_DIR, 'openspec', 'schemas', 'trinity-workflow', 'schema.yaml');
-        const schemaDest = path.join(openspecDir, 'schemas', 'trinity-workflow', 'schema.yaml');
+        // Copy selected schema
+        const schemaSrc = path.join(TEMPLATES_DIR, 'openspec', 'schemas', options.schema, 'schema.yaml');
+        const schemaDest = path.join(openspecDir, 'schemas', options.schema, 'schema.yaml');
 
         if (await fs.exists(schemaDest) && !options.force) {
-          console.log(chalk.yellow('⚠ trinity-workflow/schema.yaml already exists, use --force to overwrite'));
+          console.log(chalk.yellow(`⚠ ${options.schema}/schema.yaml already exists, use --force to overwrite`));
         } else {
           await fs.copy(schemaSrc, schemaDest);
-          console.log(chalk.green('✓ Created openspec/schemas/trinity-workflow/schema.yaml'));
+          console.log(chalk.green(`✓ Created openspec/schemas/${options.schema}/schema.yaml`));
         }
 
-        // Copy trinity-workflow templates
-        const templatesSrcDir = path.join(TEMPLATES_DIR, 'openspec', 'schemas', 'trinity-workflow', 'templates');
-        const templatesDestDir = path.join(openspecDir, 'schemas', 'trinity-workflow', 'templates');
+        // Copy schema templates if exist
+        const templatesSrcDir = path.join(TEMPLATES_DIR, 'openspec', 'schemas', options.schema, 'templates');
+        const templatesDestDir = path.join(openspecDir, 'schemas', options.schema, 'templates');
 
         if (await fs.pathExists(templatesSrcDir)) {
           await fs.copy(templatesSrcDir, templatesDestDir, { overwrite: options.force });
-          console.log(chalk.green('✓ Created openspec/schemas/trinity-workflow/templates/'));
+          console.log(chalk.green(`✓ Created openspec/schemas/${options.schema}/templates/`));
         }
 
         // Create .active file
@@ -99,11 +104,22 @@ program
       }
 
       console.log('\n' + chalk.green.bold('✅ SDD workflow initialized successfully!'));
-      console.log('\n📚 Next steps:');
-      console.log(chalk.cyan('   /sdd-new <change-name>') + '   - Start a new SDD workflow');
-      console.log(chalk.cyan('   /sdd-continue') + '         - Continue to next artifact');
-      console.log(chalk.cyan('   /sdd-apply') + '            - Execute tasks');
-      console.log(chalk.cyan('   /sdd-status') + '           - View workflow status\n');
+
+      // Show mode-specific next steps
+      if (options.schema === 'hybrid-workflow') {
+        console.log('\n📚 Hybrid Workflow Commands:');
+        console.log(chalk.cyan('   /hybrid-new <name>') + '        - Start hybrid workflow');
+        console.log(chalk.cyan('   /hybrid-explore') + '           - Explore problem space');
+        console.log(chalk.cyan('   /hybrid-ff') + '               - Fast-forward all docs');
+        console.log(chalk.cyan('   /hybrid-apply') + '             - Execute with 3-Strike');
+        console.log(chalk.cyan('   /hybrid-status') + '            - View status\n');
+      } else {
+        console.log('\n📚 Trinity Workflow Commands:');
+        console.log(chalk.cyan('   /sdd-new <change-name>') + '   - Start a new workflow');
+        console.log(chalk.cyan('   /sdd-continue') + '         - Continue to next artifact');
+        console.log(chalk.cyan('   /sdd-apply') + '            - Execute tasks');
+        console.log(chalk.cyan('   /sdd-status') + '           - View status\n');
+      }
 
     } catch (error) {
       console.error(chalk.red('\n❌ Initialization failed:'), error.message);
@@ -115,14 +131,25 @@ program
   .command('list')
   .description('List available commands and schemas')
   .action(() => {
-    console.log(chalk.bold('\n📚 Available SDD Commands:'));
+    console.log(chalk.bold('\n📚 SDD Commands (Trinity):'));
     console.log('   /sdd-new       - Start a new SDD workflow');
     console.log('   /sdd-continue  - Continue to next artifact');
     console.log('   /sdd-apply     - Execute tasks from workflow');
     console.log('   /sdd-status    - View workflow status');
 
+    console.log(chalk.bold('\n🔀 Hybrid Commands:'));
+    console.log('   /hybrid-new      - Start hybrid workflow');
+    console.log('   /hybrid-explore  - Explore problem space');
+    console.log('   /hybrid-ff       - Fast-forward all docs');
+    console.log('   /hybrid-continue - Continue to next artifact');
+    console.log('   /hybrid-apply    - Execute with 3-Strike');
+    console.log('   /hybrid-verify   - Verify implementation');
+    console.log('   /hybrid-status   - View status');
+    console.log('   /hybrid-archive  - Archive change');
+
     console.log(chalk.bold('\n📦 Available Schemas:'));
-    console.log('   trinity-workflow - 三位一体架构工作流\n');
+    console.log('   trinity-workflow  - 三位一体架构工作流');
+    console.log('   hybrid-workflow   - 融合工作流 (推荐)\n');
   });
 
 program.parse();

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "ssd-ql-workflow",
-  "version": "0.2.2",
-  "description": "SDD (Skill-Driven Development) CLI - Initialize trinity-workflow schema and commands",
+  "version": "0.4.0",
+  "description": "SDD (Skill-Driven Development) CLI - Hybrid workflow with OpenSpec flexibility + Trinity tracking",
   "type": "module",
   "bin": {
     "sdd": "bin/cli.js"

package/templates/opencode/commands/hybrid-apply.md

@@ -0,0 +1,65 @@
+---
+description: Execute tasks with 3-Strike error protocol
+---
+
+Execute implementation tasks with systematic error handling.
+
+## Input
+- Optional: Change name (defaults to active)
+- Optional: Batch number to start from
+
+## Prerequisites
+- tasks.md must exist with unchecked tasks
+- task_plan.md must exist
+
+## Execution Flow
+
+1. Load change and verify prerequisites
+
+2. Read tasks.md for task list
+
+3. Read task_plan.md for current state
+
+4. Execute tasks by batch:
+   ```markdown
+   ## Batch 1: Foundation
+   - [x] 1.1 Setup (complete)
+   - [ ] 1.2 Create model ← CURRENT
+   - [ ] 1.3 Add validation
+   ```
+
+5. After each task:
+   - Update checkbox: `- [ ]` → `- [x]`
+   - Update task_plan.md Progress
+   - Log to progress.md
+
+## 3-Strike Error Protocol
+
+When a task fails:
+
+| Attempt | Action |
+|---------|--------|
+| 1 | Diagnose and fix |
+| 2 | Try alternative approach |
+| 3 | Rethink the problem |
+| 3 fails | Escalate to user |
+
+Log all attempts in task_plan.md:
+```markdown
+| Error | Attempt | Solution |
+|-------|---------|----------|
+| TypeError: x is null | 1 | Added null check |
+| TypeError: x is null | 2 | Used optional chaining |
+| TypeError: x is null | 3 | Refactored data flow |
+```
+
+## Completion Criteria
+
+- All checkboxes in tasks.md checked
+- task_plan.md Progress: 100%
+- All verification steps passed
+
+## Output
+- All tasks implemented
+- Progress at 100%
+- Ready for /hybrid-verify

package/templates/opencode/commands/hybrid-archive.md

@@ -0,0 +1,59 @@
+---
+description: Archive completed hybrid workflow change
+---
+
+Archive a completed hybrid workflow change.
+
+## Prerequisites
+- All tasks complete (or explicitly waived)
+- Verification passed (recommended)
+
+## Steps
+
+1. Load active change
+
+2. Check artifact status:
+   - proposal.md exists?
+   - design.md exists?
+   - specs/ exists?
+   - tasks.md complete?
+
+3. Sync delta specs (if applicable):
+   - Prompt user to confirm
+   - Merge to main specs
+
+4. Archive:
+   - Move to openspec/changes/archive/YYYY-MM-DD-<name>/
+   - Update project status
+
+5. Update findings.md:
+   - Mark change as archived
+   - Add lessons learned
+
+## Output
+
+```
+You: /hybrid-archive
+
+AI: 📦 Archiving add-dark-mode...
+
+    Artifact status:
+    ✓ proposal.md
+    ✓ design.md
+    ✓ specs/ (2 files)
+    ✓ tasks.md (10/10 complete)
+
+    Delta specs: 2 files not synced
+    → Sync now? [Yes/No]
+
+You: Yes
+
+AI: ✓ Synced specs to openspec/specs/ui/dark-mode.md
+    ✓ Moved to openspec/changes/archive/2026-03-03-add-dark-mode/
+
+    Change archived successfully.
+
+    Lessons learned (add to findings.md?):
+    1. CSS variables work well for theming
+    2. Consider system preference detection for v2
+```

package/templates/opencode/commands/hybrid-continue.md

@@ -0,0 +1,55 @@
+---
+description: Continue to next artifact in hybrid workflow
+---
+
+Continue creating the next artifact in the hybrid workflow.
+
+## Steps
+
+1. Load active change
+
+2. Check current phase from task_plan.md
+
+3. Determine next artifact:
+   | Current | Next |
+   |---------|------|
+   | init | proposal |
+   | proposal | design |
+   | design | specs |
+   | specs | tasks |
+   | tasks | apply |
+
+4. Create next artifact using template
+
+5. Update task_plan.md:
+   - Phase: <new-phase>
+   - Status: complete
+   - Progress: <percentage>
+
+6. Update progress.md with action log
+
+## Output
+- Next artifact created
+- Progress updated
+- Clear next step
+
+## Example
+
+```
+You: /hybrid-continue
+
+AI: Current phase: proposal (complete)
+    Next phase: design
+
+    Creating design.md...
+
+    ✓ Architecture overview
+    ✓ Data models
+    ✓ API design
+    ✓ Component design
+    ✓ Risks & mitigations
+
+    Progress: 40%
+
+    Next: /hybrid-continue (to create specs)
+```

package/templates/opencode/commands/hybrid-explore.md

@@ -0,0 +1,39 @@
+---
+description: Explore problem space before committing to a change
+---
+
+Explore the problem space before creating a change.
+
+## When to Use
+- Requirements are unclear
+- Need to investigate options
+- Complex problem with multiple solutions
+- Performance optimization
+- Architecture decisions
+
+## Steps
+
+1. Ask user what they want to explore
+
+2. Investigate:
+   - Read relevant code
+   - Search for patterns
+   - Identify bottlenecks
+   - Research solutions
+
+3. Present findings:
+   - Problem analysis
+   - Options with pros/cons
+   - Recommendations
+   - Next steps
+
+4. Optionally create change:
+   ```
+   /hybrid-new <name> --mode structured
+   ```
+
+## Output
+- Exploration report
+- Clear understanding of problem
+- Recommended approach
+- Ready to create structured change

package/templates/opencode/commands/hybrid-ff.md

@@ -0,0 +1,60 @@
+---
+description: Fast-forward - generate all planning documents at once
+---
+
+Fast-forward through planning: generate all documents in one command.
+
+## Input
+- Optional: Change name (defaults to active change)
+
+## What it Creates
+
+1. **proposal.md** - Requirements and options
+2. **design.md** - Technical architecture
+3. **specs/**/*.md** - Feature specifications
+4. **tasks.md** - Implementation tasks
+
+## Prerequisites
+- Change must exist (use `/hybrid-new` first)
+- Tracking layer initialized
+
+## Steps
+
+1. Load active change or specified change
+
+2. Verify tracking layer exists (MUST):
+   - If missing, call planning-with-files skill
+
+3. Generate all artifacts sequentially:
+   ```
+   proposal → design → specs → tasks
+   ```
+
+4. Update task_plan.md:
+   - Phase: tasks
+   - Status: complete
+   - Progress: 80%
+
+5. Show summary and next step (/hybrid-apply)
+
+## Output
+- All planning documents created
+- Ready for implementation
+- Clear task breakdown
+
+## Example
+
+```
+You: /hybrid-ff add-dark-mode
+
+AI: 🚀 Fast-forwarding add-dark-mode...
+
+    ✓ Created proposal.md (3 options)
+    ✓ Created design.md (architecture + diagram)
+    ✓ Created specs/ui/dark-mode.md (5 scenarios)
+    ✓ Created tasks.md (12 tasks in 4 batches)
+
+    Progress: 80%
+
+    Next: /hybrid-apply
+```

package/templates/opencode/commands/hybrid-new.md

@@ -0,0 +1,52 @@
+---
+description: Start a new Hybrid workflow change
+---
+
+Start a new Hybrid workflow change with flexible mode selection.
+
+## Input
+- Change name (kebab-case) or description
+- Optional: `--mode` flag (quick | explore | structured)
+
+## Modes
+
+### Quick Mode (default)
+For clear requirements, fast execution:
+```
+/hybrid-new add-login --mode quick
+```
+Flow: init → ff → apply → archive
+
+### Explore Mode
+For unclear requirements:
+```
+/hybrid-new optimize-performance --mode explore
+```
+Flow: explore → init → continue → apply → archive
+
+### Structured Mode
+For large features:
+```
+/hybrid-new redesign-checkout --mode structured
+```
+Flow: init → proposal → design → specs → tasks → apply → archive
+
+## Steps
+
+1. If no input, ask what to build using AskUserQuestion
+
+2. Create change with hybrid-workflow schema:
+   ```bash
+   openspec new change "<name>" --schema hybrid-workflow
+   ```
+
+3. Initialize tracking layer (MUST):
+   - Use Skill tool with skill: "planning-with-files"
+   - Creates: task_plan.md, findings.md, progress.md
+
+4. Show mode-specific next steps
+
+## Output
+- Change created with hybrid-workflow schema
+- Tracking layer initialized
+- Ready for next phase based on mode

package/templates/opencode/commands/hybrid-status.md

@@ -0,0 +1,58 @@
+---
+description: Show hybrid workflow status
+---
+
+Show current hybrid workflow status and progress.
+
+## Output
+
+### Change Overview
+- Change name
+- Current phase
+- Progress percentage
+- Created date
+
+### Tracking Layer
+- task_plan.md status
+- findings.md entries
+- progress.md entries
+
+### Artifacts Status
+| Artifact | Status | Size |
+|----------|--------|------|
+| proposal.md | ✓ | 2.1k |
+| design.md | ✓ | 4.3k |
+| specs/ | ✓ | 3 files |
+| tasks.md | ✓ | 12/12 |
+
+### Recent Activity
+Last 5 entries from progress.md
+
+## Example
+
+```
+You: /hybrid-status
+
+AI: 📊 Hybrid Workflow Status
+
+    Change: add-dark-mode
+    Phase: apply
+    Progress: 85%
+
+    ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+    Artifacts
+    ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+    ✓ proposal.md    (1.2k)
+    ✓ design.md      (3.4k)
+    ✓ specs/         (2 files)
+    ✓ tasks.md       (10/12 complete)
+
+    ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+    Recent Activity
+    ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+    10:45 ✓ Batch 3.1: Added theme context
+    10:42 ✓ Batch 2.3: Updated button styles
+    10:40 ✓ Batch 2.2: Added CSS variables
+
+    Next: Continue /hybrid-apply
+```

package/templates/opencode/commands/hybrid-verify.md

@@ -0,0 +1,57 @@
+---
+description: Verify implementation matches artifacts
+---
+
+Verify implementation matches artifacts across three dimensions.
+
+## Dimensions
+
+### 1. Completeness
+- All tasks in tasks.md are checked
+- All requirements in specs have corresponding code
+- All scenarios covered
+
+### 2. Correctness
+- Implementation matches spec intent
+- Edge cases from scenarios handled
+- Error states match spec definitions
+
+### 3. Coherence
+- Design decisions reflected in code structure
+- Naming conventions consistent with design.md
+- Patterns consistent across implementation
+
+## Output Format
+
+```
+Verifying <change-name>...
+
+COMPLETENESS
+✓ All 12 tasks in tasks.md are checked
+✓ All requirements in specs have corresponding code
+⚠ Scenario "X" not tested
+
+CORRECTNESS
+✓ Implementation matches spec intent
+✓ Edge cases handled
+✓ Error states correct
+
+COHERENCE
+✓ Design decisions reflected
+✓ Naming conventions consistent
+⚠ Design mentions "X" but code does "Y"
+
+─────────────────────────────
+Critical issues: 0
+Warnings: 2
+Ready to archive: Yes (with warnings)
+
+Recommendations:
+1. Add test for scenario X
+2. Update design.md or code for Y
+```
+
+## Exit Codes
+- 0: All checks pass
+- 1: Warnings only
+- 2: Critical issues found

package/templates/opencode/commands/hybrid-workflow.md

@@ -0,0 +1,61 @@
+---
+description: Hybrid workflow - combines OpenSpec flexibility with Trinity tracking
+---
+
+Hybrid workflow combining OpenSpec flexibility with Trinity tracking.
+
+## Usage
+
+```bash
+# Quick mode (clear requirements)
+/hybrid-new <name> --mode quick
+/hybrid-ff                         # Generate all planning docs
+/hybrid-apply                      # Execute with 3-Strike protocol
+/hybrid-archive                    # Archive change
+
+# Explore mode (unclear requirements)
+/hybrid-explore                    # Explore problem space first
+/hybrid-new <name> --mode explore
+/hybrid-continue                   # Step-by-step artifact creation
+/hybrid-apply
+/hybrid-archive
+
+# Structured mode (large features)
+/hybrid-new <name> --mode structured
+/hybrid-continue                   # Create proposal
+/hybrid-continue                   # Create design
+/hybrid-continue                   # Create specs
+/hybrid-continue                   # Create tasks
+/hybrid-apply
+/hybrid-verify
+/hybrid-archive
+```
+
+## Modes
+
+| Mode | Flow | Best For |
+|------|------|----------|
+| quick | init → ff → apply → archive | Clear requirements, fast delivery |
+| explore | explore → init → continue → apply → archive | Unclear requirements, investigation |
+| structured | init → proposal → design → specs → tasks → apply → archive | Large features, formal process |
+
+## Key Features
+
+1. **Flexible Start** - Optional exploration phase
+2. **Mandatory Tracking** - Planning-with-Files initialization
+3. **Fast-Forward** - One-command planning generation
+4. **3-Strike Protocol** - Systematic error handling
+5. **Verification** - Completeness, correctness, coherence checks
+
+## Commands
+
+| Command | Description |
+|---------|-------------|
+| `/hybrid-new` | Create new hybrid workflow change |
+| `/hybrid-explore` | Explore problem space |
+| `/hybrid-ff` | Fast-forward: generate all planning docs |
+| `/hybrid-continue` | Create next artifact |
+| `/hybrid-apply` | Execute tasks with 3-Strike protocol |
+| `/hybrid-verify` | Verify implementation |
+| `/hybrid-status` | Show workflow status |
+| `/hybrid-archive` | Archive completed change |

package/templates/openspec/schemas/hybrid-workflow/schema.yaml

@@ -0,0 +1,286 @@
+name: hybrid-workflow
+version: 1
+description: |
+  融合工作流 - 结合 OpenSpec 灵活性与 Trinity 追踪能力
+
+  核心理念：
+  - 灵活启动：可选探索阶段
+  - 强制追踪：必须初始化 Planning-with-Files
+  - 快速执行：支持 ff 模式一键生成
+  - 系统化错误处理：3-Strike 协议
+
+# ================================================
+# 工作流模式
+# ================================================
+modes:
+  quick:
+    description: 快速模式 - 明确需求时使用
+    flow: ["init", "ff", "apply", "archive"]
+
+  explore:
+    description: 探索模式 - 需求不明确时使用
+    flow: ["explore", "init", "continue", "apply", "archive"]
+
+  structured:
+    description: 结构化模式 - 大型功能开发
+    flow: ["init", "proposal", "design", "specs", "tasks", "apply", "archive"]
+
+# ================================================
+# 追踪层 (强制)
+# ================================================
+tracking:
+  required: true
+  skill: planning-with-files
+  files:
+    - task_plan.md
+    - findings.md
+    - progress.md
+  instruction: |
+    [MUST] 执行任何工作流前，调用 planning-with-files 技能初始化追踪。
+
+    自动创建：
+    - task_plan.md: 阶段、目标、决策、错误日志
+    - findings.md: 技术发现、架构决策 (ADR)
+    - progress.md: 会话进度日志
+
+# ================================================
+# Workflow Artifacts 定义
+# ================================================
+artifacts:
+  # ----------------------------------------------
+  # Phase 0: 探索 (可选)
+  # ----------------------------------------------
+  - id: exploration
+    generates: exploration.md
+    description: 问题空间探索
+    optional: true
+    instruction: |
+      在需求不明确时进行探索。
+
+      探索内容：
+      1. 问题定义
+      2. 现状分析
+      3. 可能方案
+      4. 技术调研
+      5. 风险识别
+
+      输出：探索报告，指导后续 proposal
+    requires: []
+
+  # ----------------------------------------------
+  # Phase 1: 需求提案
+  # ----------------------------------------------
+  - id: proposal
+    generates: proposal.md
+    description: 需求探索与方案设计
+    template: proposal.md
+    instruction: |
+      ## 快速创建
+
+      触发词: @brainstorming 或 /brainstorming
+
+      ## 追踪初始化 [MUST]
+      执行前调用：Use the Skill tool with skill: "planning-with-files"
+
+      创建需求提案文档，聚焦于 WHY 和 WHAT。
+
+      关键点：
+      1. 问题陈述
+      2. 方案选项
+      3. 成功指标
+      4. 非目标
+      5. 影响评估
+
+      完成后更新 task_plan.md：
+      - Phase: proposal
+      - Status: complete
+      - Progress: 20%
+    requires: []
+
+  # ----------------------------------------------
+  # Phase 2: 技术设计
+  # ----------------------------------------------
+  - id: design
+    generates: design.md
+    description: 技术架构设计
+    template: design.md
+    instruction: |
+      触发词: @writing-plans 或 /writing-plans
+
+      创建技术设计文档，聚焦于 HOW。
+
+      关键点：
+      1. 架构概述
+      2. 数据模型
+      3. API 设计
+      4. 组件设计
+      5. 风险与缓解
+
+      [OPTIONAL] 架构图：Use the Skill tool with skill: "excalidraw-diagram"
+
+      完成后更新 task_plan.md：
+      - Phase: design
+      - Status: complete
+      - Progress: 40%
+
+      在 findings.md 记录架构决策（ADR）。
+    requires:
+      - proposal
+
+  # ----------------------------------------------
+  # Phase 3: 规格说明
+  # ----------------------------------------------
+  - id: specs
+    generates: specs/**/*.md
+    description: 功能规格说明
+    template: spec.md
+    instruction: |
+      创建功能规格说明文档。
+
+      使用 Gherkin 格式：
+      - Feature: 功能名称
+      - Scenario: 场景描述
+      - Given: 前置条件
+      - When: 触发动作
+      - Then: 期望结果
+
+      覆盖三种场景：
+      - Happy Path: 成功路径
+      - Edge Cases: 边界情况
+      - Error Cases: 错误处理
+
+      完成后更新 task_plan.md：
+      - Phase: specs
+      - Status: complete
+      - Progress: 60%
+    requires:
+      - design
+
+  # ----------------------------------------------
+  # Phase 4: 任务分解
+  # ----------------------------------------------
+  - id: tasks
+    generates: tasks.md
+    description: 实现任务清单
+    template: tasks.md
+    instruction: |
+      触发词: @executing-plans 或 /executing-plans
+
+      创建任务分解清单。
+
+      任务粒度要求：
+      - 每个任务 2-5 分钟可完成
+      - 每个任务有明确的验证步骤
+      - 任务按批次组织
+      - 使用 `- [ ] X.Y Task description` 格式
+
+      完成后更新 task_plan.md：
+      - Phase: tasks
+      - Status: complete
+      - Progress: 80%
+
+      准备进入 apply 阶段。
+    requires:
+      - specs
+
+# ================================================
+# Fast-Forward 模式
+# ================================================
+fastForward:
+  enabled: true
+  command: ff
+  description: 一键生成所有规划文档
+  artifacts:
+    - proposal
+    - design
+    - specs
+    - tasks
+  instruction: |
+    快速生成所有规划文档。
+
+    执行流程：
+    1. 初始化追踪层
+    2. 依次创建 proposal → design → specs → tasks
+    3. 更新进度追踪
+
+    适用场景：
+    - 需求明确
+    - 时间紧迫
+    - 中小型功能
+
+# ================================================
+# Apply 阶段配置
+# ================================================
+apply:
+  requires:
+    - tasks
+  tracks: tasks.md
+  instruction: |
+    ## 执行流程
+
+    1. 读取 tasks.md 获取任务列表
+    2. 读取 task_plan.md 获取执行状态
+    3. 按批次执行任务
+    4. 完成任务后：
+       - 更新 tasks.md 的 checkbox: `- [ ]` → `- [x]`
+       - 更新 task_plan.md 的 Progress
+       - 记录 progress.md 操作日志
+
+    ## 错误处理（3-Strike 协议）
+
+    | 尝试 | 动作 |
+    |------|------|
+    | Attempt 1 | 诊断并修复 |
+    | Attempt 2 | 尝试替代方案 |
+    | Attempt 3 | 重新思考问题 |
+    | 3次失败 | 升级给用户 |
+
+    在 task_plan.md 的错误日志中记录每次尝试。
+
+    ## 完成标准
+
+    - 所有 tasks.md 的 checkbox 已勾选
+    - task_plan.md Progress: 100%
+    - 所有验证步骤通过
+
+# ================================================
+# Verify 阶段配置
+# ================================================
+verify:
+  dimensions:
+    - completeness
+    - correctness
+    - coherence
+  instruction: |
+    验证实现与文档的一致性。
+
+    检查维度：
+    1. Completeness (完整性)
+       - 所有任务完成
+       - 所有需求实现
+       - 所有场景覆盖
+
+    2. Correctness (正确性)
+       - 实现符合规格意图
+       - 边界情况处理
+       - 错误状态匹配
+
+    3. Coherence (一致性)
+       - 设计决策体现在代码中
+       - 命名规范一致
+       - 模式使用一致
+
+# ================================================
+# Archive 阶段配置
+# ================================================
+archive:
+  syncSpecs: true
+  moveToArchive: true
+  instruction: |
+    归档完成的变更。
+
+    执行流程：
+    1. 检查所有 artifact 状态
+    2. 同步 delta specs 到 main specs
+    3. 移动到 archive 目录
+    4. 更新项目状态