v5.4.0 - CLI模板系统重组与错误处理增强

🎯 主要更新

模板系统重组

• ✨ 采用优先级前缀命名系统（01-通用, 02-专用, 03-领域特定）
• 📁 目录结构优化：bug-diagnosis 从 development 移至 analysis
• 🔗 统一引用路径格式：所有模板引用使用完整路径
• 🗑️ 移除冗余目录：implementation/, review/

模板变更详情

分析模板 (analysis/):

• 01-trace-code-execution.txt (通用-代码执行追踪)
• 01-diagnose-bug-root-cause.txt (通用-Bug根因诊断)
• 02-analyze-code-patterns.txt (专用-代码模式分析)
• 02-review-architecture.txt (专用-架构审查)
• 02-review-code-quality.txt (专用-代码质量审查)
• 03-analyze-performance.txt (领域-性能分析)
• 03-assess-security-risks.txt (领域-安全风险评估)
• 03-review-quality-standards.txt (领域-质量标准审查)

开发模板 (development/):

• 02-implement-feature.txt (专用-功能实现)
• 02-refactor-codebase.txt (专用-代码重构)
• 02-generate-tests.txt (专用-测试生成)
• 02-implement-component-ui.txt (专用-UI组件实现)
• 03-debug-runtime-issues.txt (领域-运行时调试)

规划模板 (planning/):

• 01-plan-architecture-design.txt (通用-架构规划)
• 02-breakdown-task-steps.txt (专用-任务分解)
• 02-design-component-spec.txt (专用-组件设计)
• 03-evaluate-concept-feasibility.txt (领域-概念可行性)
• 03-plan-migration-strategy.txt (领域-迁移策略)

文档模板 (memory/):

• 02-document-module-structure.txt (专用-模块文档)

错误处理增强

• 🔧 添加 Gemini 模型 404 错误回退策略
• 📝 完善错误处理文档：HTTP 429 和 HTTP 404 处理指南
• ⚡ 自动回退机制：gemini-3-pro-preview-11-2025 (404) → gemini-2.5-pro

命令文件更新

更新了 5 个命令文件中的 21 处模板引用：

• cli/mode/bug-diagnosis.md (6处)
• cli/mode/code-analysis.md (6处)
• cli/mode/plan.md (6处)
• task/execute.md (1处)
• workflow/tools/test-task-generate.md (2处)

文档改进

• 📚 更新 intelligent-tools-strategy.md：完整的模板系统指南
• 📋 新增 command-template-update-summary.md：变更详情记录

🔄 破坏性变更

无破坏性变更。所有命令引用已更新，向后兼容。

📝 提交历史

• bbddbeb: docs: 添加Gemini模型404错误回退策略
• 854464b: refactor: 重组CLI模板系统，采用优先级前缀命名

---

📦 完整变更: v5.3.0...v5.4.0

v5.3.0 - 适配 Gemini Pro 3.0 及策略优化

🚀 重大更新

✨ 新特性

适配 Gemini Pro 3.0

• 默认模型升级: 将 gemini-3-pro-preview-11-2025 设置为默认首选模型
• 模型选择优化:
  • Gemini: gemini-3-pro-preview-11-2025 (默认首选) | gemini-2.5-pro (备选) | gemini-2.5-flash (文档更新)
  • Qwen: coder-model (默认)
  • Codex: gpt-5 (默认)
• 强制模型参数: 所有CLI命令必须显式指定 -m 参数

强化模板使用规范

• 默认分析模板: RULES 字段默认包含 $(cat ~/.claude/workflows/cli-templates/prompts/analysis/pattern.txt)
• 命令示例更新: 所有示例命令添加默认模型参数和模板引用
• Core Principles 新增: 明确要求必须指定模型和使用分析模板

🔧 架构优化

精简文档结构

• 移除 Workflow Sessions (WFS) 集成: 删除所有 WFS 相关引用和说明
• 移除 SKILL Packages 集成: 简化 Memory Context，移除 SKILL 加载说明
• Memory Sources 精简: 保留 Related Tasks、Tech Stack Patterns、Cross-Module References 三类
• Best Practices 优化: 简化工作流准则，删除 WFS 和 SKILL 相关内容

命令规范强化

• 模型参数状态: 从 "optional" 改为 "REQUIRED"
• Quick Command Syntax: 添加警告提示必须指定模型和模板
• 所有示例更新: 7处示例添加默认模型参数和/或分析模板引用

📝 文档变更

Universal Template Structure

• CONTEXT 字段移除 WFS sessions 和 SKILL packages 引用
• Memory Sources 移除 workflow artifacts 路径说明
• 示例更新为通用的 commit hash 引用方式

Tool-Specific Configuration

• Gemini/Qwen 示例: 3处添加 -m gemini-3-pro-preview-11-2025
• Codex 示例: 添加 -m gpt-5 和开发模板引用
• Multi-Directory 示例: 添加模型参数和分析模板

Best Practices

• 删除 WFS sessions 和 SKILL packages 加载的详细说明（共7行）
• 简化为通用的 memory context 准则（3行）

📊 统计信息

• 文件变更: 1 file changed
• 代码精简: +94 insertions, -127 deletions
• 净减少: 33 lines (约 5.4% 精简率)
• 修复问题: 7处遗漏内容全面更新

🎯 升级指南

命令语法变更

之前:

cd [dir] && gemini -p "[prompt]"

现在:

cd [dir] && gemini -p "[prompt]" -m gemini-3-pro-preview-11-2025

RULES 字段变更

之前:

RULES: [constraints] | MODE=READ-ONLY

现在:

RULES: $(cat ~/.claude/workflows/cli-templates/prompts/analysis/pattern.txt) | [constraints] | MODE=READ-ONLY

🔗 相关提交

• afed67c: docs: 更新智能工具选择策略文档，明确规则和命令模板的使用要求
• b6b788f: 重构 intelligent-tools-strategy.md 统一提示词模板架构
• 63acd94: docs: 更新智能工具选择策略文档，增强模型选择和命令模板结构的清晰度

---

完整变更历史: v5.2.3...v5.3.0

v5.2.3 - Documentation Enhancement & Context Optimization

v5.2.3 Release Notes

📚 Documentation Enhancements

CLI Tools Strategy - Context Optimization

• --include-directories Parameter Emphasis: Comprehensive guidance on minimizing irrelevant file noise
  • Added "Multi-directory Analysis" scenario to Quick Decision Matrix
  • New "Minimize context noise" principle in Core Principles
  • Enhanced examples with detailed comments showing file scope control
  • Documented benefits: reduced token usage and improved analysis precision

Command Documentation Standardization

• Standardized command documentation format across workflow commands
• Removed emojis for cleaner, professional documentation
• Added command template specifications
• Continued workflow session and version command format standardization

Configuration & Template Updates

• Updated .gitignore to include new command template files
• Updated CLI initialization to use Write function for config creation
• Added focus_paths path clarity guidelines in task generation
• Added templates for tech stack module documentation

SKILL Package Workflow Improvements

• Enhanced SKILL.md generation descriptions and variable substitution guides
• Improved workflow-skill-memory command clarity (WFS-* session validation)
• Optimized skill-memory.md structure
• Removed SKILL package update option from workflow completion
• Updated context search strategy to clarify SKILL usage
• Enhanced path trigger strategy with intelligent skill name extraction
• Made skill_name optional with auto-detection in load-skill-memory

UI Design Workflow

• Refactored interaction patterns and field architecture

Workflow Session & Archival

• Refactored session archival and SKILL package updates
• Added workflow-skill-memory command and skill aggregation
• Removed benefits and architecture sections for clarity

🎯 Key Improvements

• Better Context Control: Clearer guidance on cd + --include-directories pattern
• Standardized Documentation: Consistent format across all commands
• Enhanced SKILL Workflow: Improved auto-detection and documentation generation
• Cleaner Examples: Added inline comments showing actual effects

📊 Statistics

• 20 commits since v5.2.2 (3eec2a5)
• Primary focus: documentation improvements and context optimization
• No breaking changes

🔧 Compatibility

• Fully backward compatible with v5.2.x
• Requires no configuration changes

v5.2.2 - Intelligent Skip Logic Enhancement

v5.2.2 Release Notes

🧠 Intelligent Skip Logic

This release adds intelligent skip logic to the /memory:skill-memory command with parameter naming corrections.

Features

• Intelligent Skip Logic: Automatically detects and skips unnecessary regeneration of SKILL packages
• Parameter Naming Correction: Fixed parameter naming consistency across commands
• Improved efficiency in documentation workflow

Commits

• feat(v5.2.2): Add intelligent skip logic to /memory:skill-memory with parameter naming correction (3eec2a5)

📊 Statistics

• Build on top of v5.2.1
• Focus: Performance optimization and consistency
• No breaking changes

🔧 Compatibility

• Fully backward compatible with v5.2.x
• Enhanced performance with intelligent caching

v5.2.1 - SKILL Workflow Improvements

v5.2.1 Release Notes

🎯 SKILL Workflow Improvements

This release focuses on enhancing the SKILL package workflow with better automation and user experience.

Features

• Improved SKILL workflow automation
• Enhanced documentation generation process
• Better integration with workflow sessions

Commits

• chore: release v5.2.1 - SKILL workflow improvements (f1c8912)

📊 Statistics

• Build on top of v5.2.0
• Focus: SKILL workflow enhancements
• No breaking changes

v5.2.0: New /memory:skill-memory Command - SKILL Package Generator

🎉 New Command: /memory:skill-memory - SKILL Package Generator

Automatically generates progressive-loading SKILL packages from project documentation with intelligent 4-phase orchestration and path mirroring.

🚀 Core Features

4-Phase Orchestrator

Fully automated workflow from documentation to SKILL package:

1. Phase 1: Prepare - Parse arguments, check existing docs, handle --regenerate
2. Phase 2: Plan - Call /memory:docs to create documentation tasks
3. Phase 3: Execute - Call /workflow:execute to generate documentation files
4. Phase 4: Index - Generate SKILL.md with progressive loading structure

Auto-Continue Mechanism: All phases run autonomously via TodoList tracking - zero manual steps required.

Progressive Loading (4 Levels)

Token-budgeted documentation access:

• Level 0: Quick Start (~2K tokens) - README only
• Level 1: Core Modules (~8K tokens) - Module READMEs
• Level 2: Complete (~25K tokens) - All modules + Architecture
• Level 3: Deep Dive (~40K tokens) - Everything + Examples

Path Mirroring Strategy

SKILL knowledge structure mirrors source code hierarchy:

Source: my_app/src/modules/auth/
  ↓
Docs: .workflow/docs/my_app/src/modules/auth/API.md
  ↓
SKILL: .claude/skills/my_app/knowledge/src/modules/auth/API.md

📋 Command Usage

/memory:skill-memory [path] [--tool <gemini|qwen|codex>] [--regenerate] [--mode <full|partial>] [--cli-execute]

Examples:

# Generate SKILL package for current project
/memory:skill-memory

# Force regeneration with Qwen
/memory:skill-memory --tool qwen --regenerate

# Partial mode (modules only)
/memory:skill-memory --mode partial

# CLI execution mode
/memory:skill-memory --cli-execute

📦 Output Structure

.claude/skills/{project_name}/
├── SKILL.md                    # Index with progressive loading levels
└── knowledge/                  # Mirrored documentation structure
    └── src/modules/auth/
        ├── API.md
        └── README.md

SKILL Package Features:

• ✅ Progressive loading (2K → 40K tokens)
• ✅ Path mirroring for easy navigation
• ✅ Intelligent capability detection
• ✅ Automatic trigger phrase generation
• ✅ Regeneration support with single flag

🔄 Enhanced /memory:docs Command

Secondary improvements to documentation workflow:

• ✅ Smart Task Grouping - ≤7 documents per task (up from 5)
• ✅ Context Sharing - Prefer grouping 2 top-level directories for shared Gemini analysis
• ✅ Batch Processing - Reduced task count through intelligent grouping
• ✅ Dual Execution Modes - Agent Mode (default) and CLI Mode (--cli-execute)
• ✅ Pre-computed Analysis - Phase 2 unified analysis eliminates redundant CLI calls
• ✅ Conflict Resolution - Automatic splitting when exceeding document limit

⚡ Performance Benefits

SKILL Package Generation:

• 📦 Zero manual steps - fully automated 4-phase workflow
• 📦 Progressive loading - load only what you need
• 📦 Auto-discovery - intelligent capability extraction

Documentation Workflow:

• ⚡ Parallel processing - multiple directory groups execute concurrently
• ⚡ Context sharing - single Gemini call per task group (2 directories)
• ⚡ Efficient analysis - one-time analysis in Phase 2, reused by all tasks
• ⚡ Predictable sizing - ≤7 docs per task ensures reliable completion

🛠️ Technical Details

New File:

• .claude/commands/memory/skill-memory.md (822 lines)

Modified Files:

• .claude/commands/memory/docs.md - Enhanced with batch processing and smart grouping
• .claude/agents/doc-generator.md - Mode-aware execution support

Multi-Tool Support:

• Gemini (default) - Comprehensive documentation
• Qwen - Architecture analysis focus
• Codex - Implementation validation

---

Previous Version: v5.1.0
Full Changelog: v5.1.0...v5.2.0

v5.1.0: Agent Architecture Consolidation

What's Changed

Agent System

• Consolidated agent architecture by replacing general-purpose agent with universal-executor
• Improved agent specialization and execution patterns

Workflow Enhancements

• Enhanced session completion command with better state management and cleanup
• Improved context-gather with advanced filtering and validation capabilities

Maintenance

• Archived legacy prompt templates for reference
• Code cleanup and organization improvements

Full Changelog: v5.0.1...v5.1.0

Release v5.0.1

What's Changed

New Features

• context-search-agent: New agent for intelligent project context collection and analysis
• Chinese Language Support: Enforced Chinese language for all user questions in brainstorm workflows

Refactoring & Improvements

• Conflict Resolution: Enhanced with interactive resolution and in-place modification suggestions
• Context Management: Streamlined context-package validation and loading in context-gather command
• Brainstorm Workflows:
  • Enhanced artifacts role selection with intelligent recommendations
  • Improved task tracking and execution phases with detailed user interaction
  • Optimized role selection with memory checks
  • Added option limits for better UX
• Agent Updates:
  • Removed code-index MCP dependencies across agents
  • Updated subagent types to context-search-agent
  • Added Windows path format guidelines
  • Enhanced execution mode and session information structure
• Documentation:
  • Refined artifacts command documentation
  • Removed redundant workflow architecture references
  • Updated file paths for clarity

Bug Fixes

• Fixed markdown formatting errors in conflict-resolution command

Changed Files

• 33 files modified
• 2,267 insertions, 1,784 deletions

Full Changelog: https://github.com/$(git remote get-url origin | sed 's/.github.com[:/](.).git/\1/')/compare/v5.0.0...v5.0.1

v5.0.0: Less is More - Simplified Architecture

🎉 Version 5.0.0: Less is More

Release Date: October 24, 2025

🌟 Core Philosophy

Version 5.0 embraces the "less is more" principle, focusing on simplification, stability, and essential features. We've removed external dependencies, streamlined workflows, and enhanced reliability by using standard, proven tools.

---

🚀 Major Changes

✅ Removed External Dependencies

• Removed MCP code-index - Replaced with standard ripgrep and find tools
• Benefits: Better stability, improved performance, easier installation
• Impact: No external service dependencies for core functionality

✅ Streamlined Workflows

• Enhanced TDD Workflow - Introduced conflict resolution mechanism
• Improved task generation - Better phase coordination and quality gates
• Simplified brainstorming - Focus on role analysis instead of complex synthesis documents

✅ Focused Architecture

• Role-based planning - Clearer separation of concerns
• Reduced abstraction - Removed over-engineered features
• Standard tools - Leveraging battle-tested utilities (ripgrep/find)

---

🔧 Breaking Changes

Removed Features

• ❌ /workflow:concept-clarify - Concept enhancement feature removed for simplification
• ❌ synthesis-specification.md workflow - Replaced with direct role analysis

Dependency Changes

• ❌ mcp__code-index dependency removed
• ✅ Standard tools (ripgrep, find) now used for file discovery
• ✅ MCP Chrome DevTools retained for UI design workflows

---

📝 Documentation Updates

All documentation has been updated to reflect v5.0 changes:

• ✅ README_CN.md - Version 5.0 highlights and philosophy
• ✅ COMMAND_REFERENCE.md - Updated command descriptions
• ✅ COMMAND_SPEC.md - Technical specifications aligned with implementation
• ✅ GETTING_STARTED.md/CN - Fixed command names and workflow descriptions
• ✅ INSTALL_CN.md - Simplified installation notes

Key Documentation Fixes

• Fixed memory command names: /update-memory-* → /memory:update-*
• Clarified test workflow execution pattern
• Removed deprecated feature references
• Updated dependency information

---

🎯 Migration Guide

For Existing Users

No immediate action required for most users. The core workflow commands remain unchanged:

/workflow:plan "your task"
/workflow:execute

Memory commands - Update your usage:

# Old (deprecated)
/update-memory-full
/update-memory-related

# New (v5.0)
/memory:update-full
/memory:update-related

Concept clarification - Feature removed:

• If you used /workflow:concept-clarify, this is no longer available
• Use /workflow:brainstorm:synthesis for role-based clarification instead

---

🔍 Under the Hood

Technical Improvements

File Discovery:

• Before: MCP code-index service
• After: Native ripgrep (rg) and find commands
• Result: Faster, more reliable, no external dependencies

Search Performance:

• Leverages ripgrep's speed for content search
• Direct filesystem access with find
• Better cross-platform compatibility

UI Design Workflows:

• Retained MCP Chrome DevTools for browser automation
• Multi-tier fallback: MCP → Playwright → Chrome → Manual
• Specialized tools for specialized tasks

---

📊 Commit History

Recent commits included in this release:

1. bd9ae8b - docs: update documentation for v5.0 release
2. da908d8 - refactor: remove MCP code-index dependency, replace with ripgrep/find
3. 3068c2c - Refactor TDD Workflow: Update phases, introduce conflict resolution
4. ee7ffda - docs: enhance workflow documentation with role analysis
5. 1f07063 - docs: update workflow plan to enhance automation
6. 57fa379 - Refactor workflow to replace synthesis-specification.md

---

🙏 Acknowledgments

Thank you to all users and contributors who provided feedback that shaped this release. Your insights on simplification and reliability were invaluable.

---

📚 Resources

• Documentation: Getting Started Guide
• Installation: Installation Guide
• Commands: Command Reference
• Issues: GitHub Issues

---

Full Changelog: v4.6.2...v5.0.0

v4.6.3 - CLI Tools & Memory System Enhancements

v4.6.3 - CLI Tools & Memory System Enhancements

🎯 Major Improvements

CLI Tools Integration

• Gemini CLI v0.11.0-nightly Migration: Upgraded to native prompt support with gemini -p syntax
• Tool Wrapper Cleanup: Removed legacy wrapper scripts for cleaner architecture
• Parameter Standardization: Unified -m model parameter placement (after prompt)

Memory System Enhancements

• Parallel Execution: Full documentation updates now use agent-based parallel processing (73% efficiency improvement)
• 3-Layer Architecture: Auto-strategy selection based on directory depth
  • Layer 3 (depth ≥3): Multi-layer strategy with full context
  • Layer 2 (depth 1-2): Single-layer aggregation strategy
  • Layer 1 (depth 0): Top-level aggregation
• Tool Fallback: Automatic 3-tier tool fallback (gemini → qwen → codex)
• Agent Task Simplification: Removed strategy judgment from agents, pre-determined by coordinator

Documentation Updates

• Intelligent Tools Strategy: Added critical directory scope rules and --include-directories usage
• Context Search Strategy: Enhanced search tool selection matrix
• Workflow Documentation: Added brainstorming workflows and UI design automation
• Getting Started Guide: Added /memory:load documentation and session auto-creation tips

📝 Detailed Changes

Memory Commands

• update_module_claude.sh: Strategy parameter support (multi-layer/single-layer)
• /memory:update-full: Agent batch processing with module-level isolation
• /memory:update-related: Enhanced context-aware updates

CLI Integration

• Model parameter placement: -m now after prompt for all tools
• Native prompt support: Direct -p flag usage
• Approval mode: --yolo for write operations

Configuration & Cleanup

• Removed deprecated tool wrapper scripts
• Cleaned up redundant tool control documentation
• Standardized command templates

🔧 Breaking Changes

CLI Command Format

Before:

gemini "prompt text" -m gemini-2.5-flash

After:

gemini -p "prompt text" -m gemini-2.5-flash

Memory Update Strategy

• Strategy now auto-determined by directory depth (no manual configuration needed)
• Agents receive pre-determined strategies from coordinator

📚 Documentation

• Updated: intelligent-tools-strategy.md - Directory scope rules
• Updated: mcp-tool-strategy.md - Tool selection matrix
• Updated: update-full.md - Agent execution flow with fallback handling
• Added: Brainstorming workflow documentation
• Added: UI design automation guides

🙏 Contributors

Thanks to all contributors who helped improve the CLI tools and memory system!

---

Full Changelog: v4.6.2...v4.6.3