ai-qakit 1.0.0

package/templates/.claude/CHANGELOG.md

@@ -0,0 +1,587 @@
+# Changelog
+
+All notable changes to the Claude Code configuration will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+---
+
+## 1.8.0 - 2025-12-22
+
+### Added
+
+- **New Spec Commands**:
+  - `/spec:deep-capture-requirement` - Comprehensive deep analysis of complex existing codebase
+  - `/spec:update-plan` - Update and refine existing implementation plan
+  - `/spec-lite:update-plan` - Lightweight version for updating spec-lite plans
+
+- **Setup Improvements**:
+  - Installation checks for Claude-Mem and MCP servers in setup scripts
+  - Pre-flight validation before dependency installation
+
+### Changed
+
+- **Spec Workflow Enhancements**:
+  - Enhanced workflow sequence with phase continuation prompts between implementation phases
+  - Updated code review process to defer review until all plan phases complete
+  - Clarified testing strategy to separate testing from implementation phase
+  - Enhanced type checking process in implementation workflow
+  - Improved requirement and plan update workflows
+
+- **Command Templates**:
+  - Updated `/spec:implement` with multi-phase continuation workflow
+  - Updated `/spec-lite:do` with user confirmation gates between phases
+  - Enhanced `/spec:brainstorm` and `/spec:capture-requirement` workflows
+  - Updated helper.py with new command mappings and TASK-ID format hints
+
+### Removed
+
+- **Hooks**:
+  - Removed `check-update.sh` hook (update notification consolidated in statusline)
+
+## 1.7.2 - 2025-12-20
+
+### Added
+
+- **Help Command** (`devkit/help.md`):
+  - All-in-one AI-DevKit usage guide with dynamic command discovery
+  - Version and changelog display via `/devkit:help version` or `/devkit:help changelog`
+  - Task-based command recommendations with output type markers for LLM guidance
+  - Complete workflow documentation for spec, spec-lite, fix, and review workflows
+
+- **Update Notification System**:
+  - Displays update banner with instructions when newer version available
+
+- **Helper Script** (`scripts/helper.py`):
+  - Powers the help command with runtime command scanning
+  - Scans `.claude/commands/` directory to build catalog dynamically
+  - Category workflows and task keyword mappings for intent detection
+
+- **Version Tracking**:
+  - `CHANGELOG.md` for Claude-specific changelog
+  - `settings.json` for Claude-specific version tracking
+
+### Changed
+
+- **Statusline Scripts** (`statusline.sh`, `statusline.ps1`):
+  - Added version update notification in statusline
+  - Shows "(New version available!)" in bright orange when newer version exists
+  - Background npm version check with 1-hour cache to avoid performance impact
+
+- **Settings**:
+  - `settings.local.json` - Added session-start hook for update notifications
+
+---
+
+## 1.7.0 - 2025-12-19
+
+### Added
+
+- **Documentation Commands**:
+  - `devkit/docs/init.md` - Initialize project documentation structure
+  - `devkit/docs/update.md` - Update existing documentation
+
+### Changed
+
+- **Command Organization**:
+  - Reorganized all commands under `commands/devkit/` folder for better namespace organization
+  - Commands now accessible via `/devkit:` prefix (e.g., `/devkit:spec:plan`, `/devkit:fix`)
+
+- **Sub-Agent Report Protocol**:
+  - Added mandatory "Phase 0: Sub-Agent Report Preference" step to all spec and spec-lite commands
+  - Orchestrator must now ask users for report preference (Inline vs Detailed) BEFORE spawning any sub-agent
+  - Commands updated: `spec/brainstorm`, `spec/capture-requirement`, `spec/cook`, `spec/plan`, `spec/implement`, `spec/test`
+  - Commands updated: `spec-lite/plan`, `spec-lite/do`, `spec-lite/test`
+  - Research command also updated with mandatory report preference gate
+
+- **Question Resolution Pattern**:
+  - Added mandatory question resolution phase to `spec/brainstorm` command
+  - Added mandatory question resolution phase to `spec/update-requirement` command
+  - Added mandatory question resolution phase to `docs/capture-knowledge` command
+  - Commands now use AskUserQuestion to resolve open questions before completing
+
+- **Agent Updates**:
+  - Updated 13 agent files with standardized report generation protocol
+  - Agents now check for `generate_report` parameter to determine output format
+  - Inline report (default): Brief summary returned to orchestrator
+  - Detailed report: Full analysis saved to task reports folder
+
+- **Statusline Scripts**:
+  - Simplified `statusline.sh` and `statusline.ps1` scripts
+  - Removed ccstatusline dependency references
+  - Cleaned up redundant setup scripts
+
+- **Documentation**:
+  - Updated 6 guide files with current command paths
+  - Updated doc-templates for improved structure
+
+### Removed
+
+- **Deprecated Scripts**:
+  - `setup/dependencies/setup-ccstatusline.ps1` - No longer needed
+  - `setup/dependencies/setup-ccstatusline.sh` - No longer needed
+- **Templates**:
+  - `doc-templates/project-roadmap.md` - Removed from template set
+
+---
+
+## 1.6.6 - 2025-12-19
+
+### Added
+
+- **Cross-Platform Statusline**:
+  - `statusline.sh` - Bash implementation for Unix/macOS with real-time session usage tracking
+  - `statusline.ps1` - PowerShell implementation for Windows environments
+  - `statusline.cjs` - Cross-platform Node.js runner that auto-detects OS and delegates to appropriate script
+  - Displays git branch, Claude model info, DevKit version, and session reset timers
+  - Color-coded usage indicators (green <50%, yellow 50-79%, red ≥80%)
+  - Auto-installs `ccusage` for token tracking if missing
+- **Settings Configuration**:
+  - `.claude/settings.json` - Project settings with AI_DEVKIT_VERSION and statusline configuration
+
+### Changed
+
+- **Commands**:
+  - Updated `update-changelog.md` to always update `.claude/settings.json` version (not conditional)
+  - Updated `commands/fix/hard.md` - Simplified project management workflow
+  - Updated `commands/spec-lite/test.md` - Minor improvements
+
+---
+
+## 1.6.5 - 2025-12-19
+
+### Added
+
+- **Auto-Setup Scripts**:
+  - Complete setup system with `setup.sh` (macOS/Linux) and `setup.ps1` (Windows) entry points
+  - Cross-platform setup scripts for 3 Claude Code enhancements:
+    - `setup-claude-mem` - Persistent memory compression system with web viewer
+    - `setup-ccstatusline` - Customizable status line formatter
+    - `setup-mcp-servers` - MCP servers (context7 for docs, sequential-thinking for analysis)
+  - All scripts include `--diagnose`, `--fix`, `--status` flags for troubleshooting
+  - Organized under `setup/dependencies/`
+
+### Changed
+
+- **Settings Documentation**:
+  - Enhanced hook documentation in `settings.local.README.md`
+  - Clarified 4-step delegation sequence for `executable_force-agent-skills-eval.sh`
+  - Expanded modularization guidelines for `dev-rules-reminder.cjs`
+  - Added cross-platform details for `scout-block.cjs` (Windows/Unix/WSL)
+
+- **Commands**:
+  - Moved `create-test.md` from `frontend/` subfolder to root commands directory
+
+### Removed
+
+- Duplicate `requirements.txt` from `scripts/` (consolidated to root)
+
+---
+
+## 1.6.4 - 2025-12-19
+
+### Changed
+
+- **Orchestration Protocol (CLAUDE.md)**:
+  - Added new **Codebase Search Protocol** section mandating `scout` sub-agent for codebase searches
+  - Updated anti-patterns to explicitly recommend `scout` sub-agent over grep/glob for multi-file searches
+  - Provides clear guidance on when to use scout: finding files, functions, patterns, understanding implementations, locating configurations
+
+---
+
+## 1.6.3 - 2025-12-18
+
+### Added
+
+- **Structured Context Protocol**:
+  - New protocol for passing task metadata (task-id, platform, task-type, files, technologies) from main agent to sub-agents
+  - Enables deterministic skill detection based on context signals
+  - Context-to-skill mapping (e.g., `platform: frontend` + `task-type: ui` → `frontend-development`, `ui-ux-pro-max`)
+
+- **Sub-Agent Report Generation Protocol**:
+  - Orchestrator must ask user about report preference before spawning report-capable agents
+  - `generate_report: true/false` parameter for sub-agent prompts
+
+- **Skill Selection Matrix**:
+  - Added to `coder` agent for automatic skill activation based on task type
+  - Covers UI, database, API, TypeScript, Python, security, performance, infrastructure, and mobile development
+
+### Changed
+
+- **Agent Templates**:
+  - Updated all 13 agent templates with refined report generation protocol
+  - Added skill usage reminders to sub-agent invocations across all commands
+  - Removed `AskUserQuestion` tool from `coder` agent (sub-agents cannot ask users directly)
+
+- **Requirement File Naming**:
+  - Standardized requirement file naming with numeric prefixes for consistent ordering
+  - Updated spec commands to use new naming convention (e.g., `01-functional-requirements.md`)
+
+- **Command Templates**:
+  - Updated 38 command templates with report protocol reminders
+  - Enhanced `spec-lite/do.md`, `spec/implement.md`, `spec/cook.md` with Structured Context Protocol
+  - Added skill usage reminders to `tester`, `code-reviewer`, `scout`, and `planner` sub-agent invocations
+
+---
+
+## 1.6.2 - 2025-12-18
+
+### Added
+
+- **New Testing Commands**:
+  - `/spec:test` - Comprehensive file-by-file testing with 100% coverage target
+  - `/spec-lite:test` - Streamlined testing for lite tasks
+
+- **Tester Agent Per-File Mode**:
+  - Per-file testing workflow with coverage commands per language
+  - Per-file output format for detailed reporting
+
+### Changed
+
+- **Testing Separation from Implementation**:
+  - `/spec:implement` now focuses solely on coding (testing via `/spec:test`)
+  - `/spec-lite:do` no longer includes testing (testing via `/spec-lite:test`)
+  - `/spec:plan` no longer asks about test planning preferences
+
+- **Mandatory User Approval Gate**:
+  - `/spec:implement` Step 4 now requires explicit user approval before finalizing
+  - `/spec-lite:do` Step 4 now requires `AskUserQuestion` for user approval
+  - Implementation cannot proceed without user confirmation
+
+- **Tester Agent Enhancements**:
+  - Added "CRITICAL MANDATE" - tester MUST fix failing tests until all pass
+  - Introduced Fix Loop with strategies by error type
+  - Added repository files (`*.repository.ts`) to non-testable filter
+  - Forbidden actions: commenting out tests, using skip/xtest, changing assertions
+
+- **Update-Requirement Command Simplification**:
+  - Now lightweight consolidation only (reads existing files, outputs `functional.md` + `notes.md`)
+  - Removed codebase analysis and Ultrathink directive
+  - Removed scope clarification questions
+
+### Removed
+
+- **docs-manager Sub-Agent**:
+  - Removed from `/spec:implement` workflow
+  - Removed from `/spec-lite:do` workflow
+  - Documentation management now handled separately
+
+---
+
+## 1.6.1 - 2025-12-18
+
+### Added
+
+- **Agent Enhancements**:
+  - Added explicit `tools` field to multiple agents (brainstormer, tester) for clearer capability definition
+  - Added user confirmation prompt before report generation in brainstormer and tester agents (for token consumption efficiency)
+  - Added File Scope filtering to `tester` agent - now excludes migrations, docs, config, constants, types, seeds, mocks, and build files from testing
+
+### Changed
+
+- **Spec Commands**:
+  - `spec:implement` - Testing step now marked as OPTIONAL, only triggers after all plan phases complete
+  - `spec:implement` - Cleanup step only removes active-task file on final phase
+  - `spec:implement` - Updated development rules reference to include both project-specific and general rules
+- **README Documentation**:
+  - Updated skills table: added `code-reviewer`, `database-admin`, `aesthetic`, `chrome-devtools`, `docs-discovery`
+  - Updated document-skills description for clarity
+
+### Removed
+
+- **Commands**:
+  - `analysis:optimize-api-performance` - Removed from Claude commands
+- **Hooks**:
+  - `post-task-completion.sh` - Removed PostToolUse hook for Task completion
+  - Removed PostToolUse hook configuration from settings.local.json
+
+---
+
+## 1.6.0 - 2025-12-18
+
+### Added
+
+- **Spec Lite Commands** (Lightweight Task Management):
+  - `spec-lite:task` - Create tasks with inline requirements
+  - `spec-lite:plan` - Quick planning for small tasks
+  - `spec-lite:do` - Execute lightweight tasks
+
+- **Configuration & Setup**:
+  - `config:bootstrap` - New bootstrap command for step-by-step project initialization
+  - `.claude/Makefile` - Build automation for Claude templates
+  - `.claude/.scoutignore` - File exclusion patterns for scout operations
+  - Python scripts for catalog generation in `scripts/`:
+    - `scan_commands.py`, `scan_skills.py`, `generate_catalogs.py`
+    - `commands_data.yaml`, `skills_data.yaml` data files
+
+- **New Skills**:
+  - **chrome-devtools** - Browser automation via Puppeteer CLI scripts with persistent sessions
+  - **aesthetic** - Create beautiful interfaces following proven design principles
+    - skill-creator moved into aesthetic namespace
+
+- **Documentation**:
+  - `guide/MCP_GUIDE.md` - Model Context Protocol setup and recommended servers
+  - `skills/agent_skills_spec.md` - Agent skills specification
+
+### Changed
+
+- **Commands**:
+  - `spec:capture-requirement` - Added Ultrathink directive and active-task saving support
+  - `spec:plan` - Added Ultrathink directive and active-task saving support
+  - `spec:implement` - Enhanced with active-task saving, auto-commit removed, cleanup step added
+  - `config:init-backend` - Enhanced initialization process
+  - `scout` - Updated to support .scoutignore patterns
+  - `skill:optimize` - Improvements to skill optimization workflow
+  - `review:codebase` - Enhanced codebase review process
+
+- **Agents** (Quality improvements across all):
+  - brainstormer, code-reviewer, code-simplifier, coder
+  - database-admin, debugger, docs-manager, planner
+  - project-manager, researcher, scout, tester, ui-ux-designer
+
+- **Skills**:
+  - ai-multimodal - Updated setup scripts and batch processing
+  - code-review - Enhanced reception, requesting, and verification guides
+  - document-skills - Improved DOCX and PPTX processing scripts
+  - sequential-thinking - Updated thought formatting and processing
+  - ui-ux-pro-max - General improvements
+
+- **Hooks**:
+  - `dev-rules-reminder` - Substantial refactoring
+  - `scout-block` - Updated to use .scoutignore, reorganized into subdirectory
+
+- **Templates & Documentation**:
+  - `CLAUDE.md` - Enhanced orchestration rules and command quality guidelines
+  - Quick start guides expanded with Spec Lite command documentation
+  - README files updated
+  - `doc-templates/codebase-summary.md` - Template improvements
+  - `workflows/development-rules.md` - Updated development standards
+  - `workflows/documentation-management.md` - Enhanced documentation practices
+
+### Fixed
+
+- **scout-block hook** - Now correctly uses `.scoutignore` file for exclusion patterns
+
+### Removed
+
+- **Deprecated Commands**:
+  - `config:init-full.md` - Removed in favor of bootstrap command
+  - `config:sync-ai-context.md` - Deprecated sync command
+  - `config:sync-incremental.md` - Deprecated incremental sync command
+
+---
+
+## 1.5.2 - 2025-12-17
+
+### Added
+- **Agents**:
+  - `database-admin` - Senior database administrator agent for PostgreSQL/MySQL/MongoDB
+
+- **Hooks**:
+  - `dev-rules-reminder` - PreToolUse hook injecting development rules and modularization reminders (JS/CJS variants)
+  - `scout-block` - Cross-platform hook blocking access to heavy directories with configurable `.scoutignore` patterns
+
+- **Skills Scripts**:
+  - `ai-multimodal/scripts/check_setup.py` - Interactive setup verification
+
+### Changed
+- **Agents**:
+  - All agents now include standardized Report Output sections with location resolution
+  - `code-reviewer` - Added report output guidelines and location resolution
+  - `debugger` - Added report output specifications and active task state management
+  - `docs-manager` - Changed model to haiku for efficiency
+  - `planner` - Changed model to opus; added active task state management
+  - `project-manager` - Added report output specifications
+  - `researcher` - Refactored for conciseness; changed model to haiku
+  - `scout` - Updated configurations and structure
+  - `tester` - Enhanced with report output specifications
+  - `ui-ux-designer` - Enhanced with report output specifications
+
+- **Commands**:
+  - `/spec:cook` - Enhanced with explicit `AskUserQuestion` tool usage
+  - `/spec:implement` - Major refactoring: new 6-step workflow with TodoWrite integration
+  - `/spec:plan` - Added `planning` skill activation requirement
+
+- **Skills**:
+  - `ai-multimodal/SKILL.md` - Refactored; added Imagen 4, Veo 3.x, Gemini 3 Pro Image support
+  - `planning/SKILL.md` - Enhanced content
+
+- **Workflows**:
+  - `documentation-management.md` - Major enhancements
+
+- **Configuration**:
+  - `settings.local.json` - Added permissions to restrict certain Git commands
+
+### Removed
+- **Hooks**:
+  - `modularization-hook.js` - Replaced by `dev-rules-reminder` hook
+
+- **Development Rules**:
+  - `rules/09-version-control.md` - Version Control & Git rules module removed
+
+---
+
+## 1.5.1 - 2025-12-16
+
+### Added
+- **Agents**:
+  - `code-reviewer` - Comprehensive code quality assessment and best practices enforcement agent
+
+### Changed
+- **Commands**:
+  - `/spec:plan` - Added API development approach detection (schema-first vs code-first)
+  - `/spec:implement` - Enhanced to respect plan metadata
+
+---
+
+## 1.5.0 - 2025-12-15
+
+### Added
+- **Documentation**:
+  - `TROUBLESHOOTING.md` - Comprehensive troubleshooting guide
+
+- **Hooks**:
+  - `post-task-completion.sh` - PostToolUse hook triggered after Task tool completion
+
+- **Configuration Documentation**:
+  - `settings.local.README.md` - Comprehensive guide for hooks configuration
+
+- **Skills References**:
+  - `code-review/references/human-code-review-practices.md`
+
+- **Commands**:
+  - `/analyze` - Multi-domain code analysis command
+
+- **Agents**:
+  - `performance-engineer` - Measurement-driven performance optimization specialist
+  - `quality-engineer` - Comprehensive testing strategist
+  - `storage-path-convention.md` - Agent documentation defining standardized file storage structure
+
+- **JavaScript Pro Skill**:
+  - Complete reference set with modern ES2015+ JavaScript guidance
+
+- **TypeScript Pro Skill**:
+  - Advanced TypeScript type system mastery
+
+- **Python Pro Skill**:
+  - Modern Python development with type hints, async patterns
+
+### Changed
+- **Skills Organization**:
+  - Moved Python Pro skill for consistent template structure
+
+- **Agents**:
+  - `security-agent.md` - Added context framework note
+  - Standardized storage path convention across all agents
+  - Updated task directory structure to `.claude/tasks/[TASK-ID]/[PHASE]/`
+
+- **Skills**:
+  - `code-review` - Consolidated code-review-excellence skill content
+
+- **Hooks Configuration**:
+  - `settings.local.json` - Added PostToolUse hook for Task tool
+
+### Removed
+- **Skills**:
+  - `code-review-excellence` - Consolidated into `code-review` skill
+
+---
+
+## 1.4.1 - 2025-12-15
+
+### Changed
+- **Frontend Development Skill**:
+  - Refocused from multi-framework to React 18+ specialization
+  - Adopted Feature Sliced Design (FSD) as standard architecture pattern
+  - Reorganized from `resources/` to `references/` folder structure
+
+### Added
+- **Frontend Development References**:
+  - Complete reference set (13 files) for React 18+ development
+
+### Removed
+- **Legacy Resources**:
+  - Consolidated and replaced with specialized references
+
+---
+
+## 1.4.0 - 2025-12-15
+
+### Added
+- **Commands**:
+  - `/review:code` command - Validates code against development rules
+
+- **Development Rules** (9-file modular system):
+  - `01-core-principles.md` through `index.md`
+
+### Changed
+- **Commands**:
+  - `update-readme` command now explicitly prohibits adding numeric counts
+  - `spec/implement.md` enhanced with code review phase integration
+  - `spec/plan.md` enhanced workflow structure
+  - `docs/init.md` updated documentation initialization
+
+- **Agents**:
+  - `docs-manager.md` enhanced with improved capabilities
+
+- **Templates**:
+  - `CLAUDE.md` updated orchestration and workflow guidance
+
+### Removed
+- `common-development-rules.md` - Replaced by modular 9-file rule system
+
+---
+
+## 1.3.0 - 2025-12-14
+
+### Changed
+- **Commands**:
+  - `/research` - Consolidated from `/research:deep` to root-level command
+
+### Removed
+- **Commands**:
+  - `/fix:logs` - Consolidated into main debugging workflow
+  - `/fix:test` - Consolidated into main debugging workflow
+  - `/fix:types` - Consolidated into main debugging workflow
+  - `/review:deep` - Consolidated into `/review:codebase`
+  - `/security:audit` - Moved to skill-based approach
+  - `/skill:fix-logs` - Deprecated
+
+---
+
+## 1.2.0 - 2025-12-13
+
+### Added
+- **New Skills** (7 skills):
+  - `databases`, `docs-discovery`, `frontend-development`, `infra-engineer`
+  - `mobile-development`, `prompt-enhancer`, `ui-ux-pro-max`
+- **Claude Hooks**:
+  - `force-agent-skills-eval.sh` - Mandatory agent/skill evaluation before task execution
+  - `modularization-hook.js` - File size monitoring and modularization suggestions
+
+### Changed
+- **Ultrathink Extended Thinking**:
+  - Added `Ultrathink` directive to `brainstormer` agent
+  - Added `Ultrathink` directive to spec commands
+
+---
+
+## 1.0.0 - 2025-12-09
+
+### Added
+- **Frontend Commands**: `/create-component`, `/create-page`, `/create-form`, `/create-table`, `/create-api-integration`, `/create-storybook`
+- **Spec-Driven Development Commands**: `/spec:task`, `/spec:brainstorm`, `/spec:capture-requirement`, `/spec:cook`, `/spec:plan`, `/spec:implement`
+- **Multi-Agent Commands**: `/implement`, `/plan`, `/refactor`, `/solve`, `/security:audit`, `/review:deep`, `/optimize:db`, `/research:deep`, `/update-changelog`
+- **New Agents** (18 total): brainstormer, debugger, planner, project-manager, scout, tester, ui-ux-designer, docs-manager
+- **New Skills** (15 total): ai-multimodal, api-design-principles, backend-development, code-review-excellence, debugging, docs-seeker, planning, problem-solving, repomix, research, sequential-thinking, document-skills
+- **Documentation Templates**: code-standards.md, codebase-summary.md, project-overview-pdr.md, project-roadmap.md, system-architecture.md, technical-documents.md
+- **Workflows**: development-rules, documentation-management, orchestration-protocol
+
+---
+
+## 0.1.0 - 2025-12-02
+
+### Added
+- Initial backend devkit for Claude Code
+- Base command structure and agent templates
+- Core skills and workflow definitions