@wipal/agent-team 1.0.4 → 1.1.0

package/README.md

@@ -6,11 +6,11 @@
 
 ## Overview
 
-Agent Team Template là hệ thống để:
-- Tạo team với nhiều AI agents chuyên môn
-- Mỗi agent có skills, knowledge base riêng
-- Tự học và cải thiện qua thời gian
-- Setup nhanh cho mỗi dự án
+Agent Team Template is a system for:
+- Creating teams with multiple specialized AI agents
+- Each agent has their own skills and knowledge base
+- Self-learning and improvement over time
+- Quick setup for each project
 
 ## Quick Start
 
@@ -30,10 +30,11 @@ npx @wipal/agent-team ui
 | `init` | Initialize .claude/ structure |
 | `add <name> <role>` | Add new agent |
 | `list` | List all agents |
-| `switch <name>` | Show switch instructions |
 | `remove <name>` | Remove agent |
 | `ui` | Start web dashboard |
 
+> **Note:** Use Claude Code's built-in `/agents` command to switch between agents.
+
 ### Add Command Options
 
 ```bash
@@ -62,6 +63,9 @@ npx @wipal/agent-team add payment-fe dev-fe --framework react --styling tailwind
 # Backend agent with NestJS + Prisma
 npx @wipal/agent-team add auth-api dev-be --framework nestjs --orm prisma
 
+# Designer agent
+npx @wipal/agent-team add ui-designer designer
+
 # Tech Lead (inherits dev-fe + dev-be skills)
 npx @wipal/agent-team add lead tech-lead
 
@@ -86,28 +90,31 @@ Features:
 
 ## Available Roles
 
-| Role | Description | Skills Count |
-|------|-------------|--------------|
-| **dev-fe** | Frontend Developer | 8 |
-| **dev-be** | Backend Developer | 8 |
-| **sa** | Solution Architect | 9 |
-| **tech-lead** | Tech Lead (inherits dev-fe + dev-be) | 17 |
-| **devops** | DevOps Engineer | 8 |
-| **pm** | Product Manager | 8 |
-| **qa** | QA Engineer | 7 |
+| Role | Description | Tool Profile | Skills Count |
+|------|-------------|--------------|--------------|
+| **dev-fe** | Frontend Developer | coding | 11 |
+| **dev-be** | Backend Developer | coding | 10 |
+| **designer** | UI/UX Designer (web/mobile) | minimal | 10 |
+| **sa** | Solution Architect | full | 13 |
+| **tech-lead** | Tech Lead (inherits dev-fe + dev-be) | orchestration | 20 |
+| **devops** | DevOps Engineer | coding | 10 |
+| **pm** | Product Manager | research | 11 |
+| **ba** | Business Analyst | research | 11 |
+| **qa** | QA Engineer | coding | 9 |
 
 ## Skills Overview
 
-### Total: 37 Skills across 8 categories
+### Total: 50 Skills across 9 categories
 
 | Category | Count | Description |
 |----------|-------|-------------|
-| **Core** | 3 | Universal skills (code-review, git-automation, retrospect) |
-| **Frontend** | 5 | React/Vue/UI development |
+| **Core** | 5 | Universal skills (code-review, git-automation, retrospect, agent-creation, sequential-thinking) |
+| **Frontend** | 6 | React/Vue/UI development, ui-ux-pro-max |
 | **Backend** | 5 | API/Database/Security |
-| **Architecture** | 6 | System design, ADRs |
+| **Design** | 5 | UI design, design systems, mockups, responsive, HTML/CSS output |
+| **Architecture** | 8 | System design, ADRs, C4, Mermaid diagrams |
 | **DevOps** | 5 | CI/CD, Docker, Terraform |
-| **Product** | 5 | Requirements, Sprints |
+| **Product** | 6 | Requirements, Sprints, Requirements clarity |
 | **Quality** | 4 | Testing, QA |
 | **Leadership** | 4 | Tech Lead skills |
 
@@ -117,26 +124,25 @@ Features:
 .claude/skills/
 ├── SKILL-INDEX.md              # Central hub with dependency graph
 │
-├── core/                       # Universal skills (3)
+├── core/                       # Universal skills (5)
 │   ├── code-review/
 │   ├── git-automation/
-│   └── retrospect-work/
+│   ├── retrospect-work/
+│   ├── agent-creation/
+│   └── sequential-thinking/
 │
-├── domain/                     # Domain-specific (31)
-│   ├── frontend/               # 5 skills
+├── domain/                     # Domain-specific (39)
+│   ├── frontend/               # 6 skills
 │   ├── backend/                # 5 skills
-│   ├── architecture/           # 6 skills
+│   ├── design/                 # 5 skills
+│   ├── architecture/           # 8 skills
 │   ├── devops/                 # 5 skills
-│   ├── product/                # 5 skills
+│   ├── product/                # 6 skills
 │   └── quality/                # 4 skills
 │
 ├── leadership/                 # Tech Lead skills (4)
-│   ├── code-review-advanced/
-│   ├── technical-decision/
-│   ├── mentoring/
-│   └── technical-debt/
 │
-└── community/                  # Community contributions
+└── community/                  # Community contributions (1)
 ```
 
 ### Skill Dependency Mechanism
@@ -151,13 +157,35 @@ description: What and WHEN to use this skill
 category: domain-name
 tags: [tag1, tag2]
 
-# Dependency mechanism - "Skill dùng Skill"
+# Dependency mechanism
 depends_on: []           # Hard dependencies (must load)
 recommends: []           # Soft dependencies (suggest in context)
 used_by: []              # Reverse reference (for discovery)
 ---
 ```
 
+## Memory System
+
+Each agent has a memory system for cross-session learning. This is **complementary** to Claude Code's built-in auto-memory.
+
+### Two Memory Systems
+
+| System | Location | Purpose |
+|--------|----------|---------|
+| Claude Code Auto-Memory | `~/.claude/projects/<project>/memory/` | Project-wide context (automatic) |
+| Agent Memory | `.claude/agent-memory/<name>/` | Per-agent specialized context |
+
+### Agent Memory Files
+
+| File | Purpose | Auto-loaded |
+|------|---------|-------------|
+| `MEMORY.md` | Session memory (what was discussed/done) | Yes (first 200 lines) |
+| `knowledge.md` | Long-term knowledge (skills, patterns, case studies) | No (manual reference) |
+
+Update memory via:
+- `/learn` skill - Capture patterns quickly
+- `/retrospect-work` skill - Session retrospectives
+
 ## Documentation
 
 | Doc | Description |
@@ -198,6 +226,51 @@ used_by: []              # Reverse reference (for discovery)
 - Pattern promotion (Lesson → Pattern → Skill → Rule)
 - Cross-session memory
 
+### 6. MCP Integration
+- **Context7**: Documentation lookup
+- **GitHub**: Repository operations
+- **Sequential Thinking**: Complex problem solving
+- **DrawIO**: Diagram and mockup creation
+
+### 7. Tool Profiles (v2.1)
+
+Predefined tool sets for different role types (inspired by OpenFang):
+
+| Profile | Tools | Used By |
+|---------|-------|---------|
+| **minimal** | read, write, edit | designer |
+| **coding** | read, write, edit, bash, grep, glob + filesystem, github | dev-fe, dev-be, devops, qa |
+| **research** | read, web_search, web_fetch + context7, docs-seeker | pm, ba |
+| **orchestration** | coding tools + task + context7 | tech-lead |
+| **full** | all tools + all MCP servers | sa |
+
+### 8. Capability-Based Security (v2.1)
+
+Fine-grained permissions in role rules:
+
+```markdown
+## Tool Access Control
+
+### File Operations
+- READ: src/**, public/**
+- WRITE: src/**
+- DENY: .env*, secrets/**
+
+### Shell Commands
+- ALLOW: npm run *, git *
+- DENY: rm -rf, sudo *
+```
+
+### 9. Operational Phases (v2.1)
+
+Multi-phase methodology for consistent workflows:
+
+1. **ANALYZE** - Read context, understand requirements
+2. **PLAN** - Design approach, consider alternatives
+3. **IMPLEMENT** - Write clean code, follow conventions
+4. **VERIFY** - Run tests, validate output
+5. **DOCUMENT** - Update docs, capture lessons
+
 ## Configuration
 
 ### Per-Project Role Override
@@ -219,13 +292,14 @@ roles:
 
 ## Research Sources
 
-Skills và architecture được xây dựng dựa trên research từ:
+Skills and architecture are built based on research from:
 
 1. **Agent-Skills-for-Context-Engineering** - Context engineering patterns
 2. **AI-Research-SKILLs** - 85 skills, 21 categories structure
 3. **everything-claude-code** - Hooks, commands, instinct system
 4. **BMAD-METHOD** - Agent YAML format, sharded workflows
 5. **spec-kit** - Extension system, multi-agent support
+6. **OpenFang** - Tool profiles, capability-based security, orchestrator patterns (v2.1)
 
 ## License
 

package/README.npm.md

@@ -0,0 +1,252 @@
+# @wipal/agent-team
+
+> **Template for creating AI agent teams with specialized roles, skills, and workflows**
+
+---
+
+## Overview
+
+Agent Team Template is a system for:
+- Creating teams with multiple specialized AI agents
+- Each agent has its own skills and knowledge base
+- Self-learning and improvement over time
+- Quick setup for each project
+
+## Quick Start
+
+```bash
+# Via npx (recommended)
+npx @wipal/agent-team init
+npx @wipal/agent-team add payment-fe dev-fe --framework react --styling tailwind
+
+# Or use UI dashboard
+npx @wipal/agent-team ui
+```
+
+## CLI Commands
+
+| Command | Description |
+|---------|-------------|
+| `init` | Initialize .claude/ structure |
+| `add <name> <role>` | Add new agent |
+| `list` | List all agents |
+| `switch <name>` | Show switch instructions |
+| `remove <name>` | Remove agent |
+| `projects` | Manage registered projects |
+| `setup-hooks` | Configure Claude Code hooks |
+| `ui` | Start web dashboard |
+
+### Add Command Options
+
+```bash
+npx @wipal/agent-team add <name> <role> [options]
+
+# Frontend options
+--framework <name>      # react, vue, angular, svelte
+--styling <name>        # tailwind, css-modules, styled-components
+--state <name>          # zustand, redux, recoil
+
+# Backend options
+--database <name>       # postgresql, mysql, mongodb
+--orm <name>            # prisma, typeorm, sequelize
+
+# Common options
+--testing <name>        # vitest, jest, cypress
+-i, --interactive       # Interactive mode
+```
+
+### Examples
+
+```bash
+# Frontend agent with React + Tailwind
+npx @wipal/agent-team add payment-fe dev-fe --framework react --styling tailwind
+
+# Backend agent with NestJS + Prisma
+npx @wipal/agent-team add auth-api dev-be --framework nestjs --orm prisma
+
+# Tech Lead (inherits dev-fe + dev-be skills)
+npx @wipal/agent-team add lead tech-lead
+
+# Interactive mode
+npx @wipal/agent-team add
+```
+
+## Web Dashboard
+
+```bash
+npx @wipal/agent-team ui
+
+# Opens at http://localhost:3456
+```
+
+Features:
+
+- Agent management (CRUD)
+- Role configuration (YAML editor)
+- Skill browser
+- Variant picker
+- Dashboard stats
+
+## Available Roles
+
+| Role | Description | Skills Count |
+|------|-------------|--------------|
+| **dev-fe** | Frontend Developer | 8 |
+| **dev-be** | Backend Developer | 8 |
+| **sa** | Solution Architect | 9 |
+| **tech-lead** | Tech Lead (inherits dev-fe + dev-be) | 17 |
+| **devops** | DevOps Engineer | 8 |
+| **pm** | Product Manager | 8 |
+| **qa** | QA Engineer | 7 |
+
+## Skills Overview
+
+### Total: 37 Skills across 8 categories
+
+| Category | Count | Description |
+|----------|-------|-------------|
+| **Core** | 3 | Universal skills (code-review, git-automation, retrospect) |
+| **Frontend** | 5 | React/Vue/UI development |
+| **Backend** | 5 | API/Database/Security |
+| **Architecture** | 6 | System design, ADRs |
+| **DevOps** | 5 | CI/CD, Docker, Terraform |
+| **Product** | 5 | Requirements, Sprints |
+| **Quality** | 4 | Testing, QA |
+| **Leadership** | 4 | Tech Lead skills |
+
+### Skills Structure
+
+```
+.claude/skills/
+├── SKILL-INDEX.md              # Central hub with dependency graph
+│
+├── core/                       # Universal skills (3)
+│   ├── code-review/
+│   ├── git-automation/
+│   └── retrospect-work/
+│
+├── domain/                     # Domain-specific (31)
+│   ├── frontend/               # 5 skills
+│   ├── backend/                # 5 skills
+│   ├── architecture/           # 6 skills
+│   ├── devops/                 # 5 skills
+│   ├── product/                # 5 skills
+│   └── quality/                # 4 skills
+│
+├── leadership/                 # Tech Lead skills (4)
+│   ├── code-review-advanced/
+│   ├── technical-decision/
+│   ├── mentoring/
+│   └── technical-debt/
+│
+└── community/                  # Community contributions
+```
+
+### Skill Dependency Mechanism
+
+Each skill has a YAML frontmatter with dependencies:
+
+```yaml
+---
+name: skill-name
+version: 1.0.0
+description: What and WHEN to use this skill
+category: domain-name
+tags: [tag1, tag2]
+
+# Dependency mechanism - "Skill using Skill"
+depends_on: []           # Hard dependencies (must load)
+recommends: []           # Soft dependencies (suggest in context)
+used_by: []              # Reverse reference (for discovery)
+---
+```
+
+## Documentation
+
+| Doc | Description |
+|-----|-------------|
+| [TUTORIAL.md](TUTORIAL.md) | Step-by-step guide |
+| [Architecture](docs/01-architecture.md) | System design and structure |
+| [Setup Guide](docs/02-setup-guide.md) | How to setup agents |
+| [Skills Guide](docs/03-skills-guide.md) | Managing skills |
+| [Workflows](docs/04-workflows.md) | Workflow orchestration |
+| [Model Strategy](docs/05-model-strategy.md) | Model selection by task |
+| [Extend Guide](docs/06-extend-guide.md) | Creating new agents/skills |
+| [Quick Reference](docs/07-quick-reference.md) | Cheat sheet |
+
+## Key Features
+
+### 1. Base + Variants Architecture
+
+- Base role provides common skills/MCPs
+- Variants add tech-specific capabilities
+- Agents can combine multiple variants
+
+### 2. Skill Dependency System
+
+- Skills can depend on other skills
+- Progressive disclosure (metadata → instructions → resources)
+- Central hub for navigation
+
+### 3. Role Inheritance (v2)
+
+- Tech Lead inherits dev-fe + dev-be skills
+- Configurable via `config/roles.yaml`
+- Per-project override in `.claude/roles.yaml`
+
+### 4. Rules System
+
+- General rules for all agents
+- Role-specific rules
+- Lessons learned from corrections
+
+### 5. Self-Improvement
+
+- Capture lessons from corrections
+- Pattern promotion (Lesson → Pattern → Skill → Rule)
+- Cross-session memory
+
+### 6. Project Management
+
+- Global registry tracks all initialized projects
+- `projects --list` to see all projects
+- `projects --prune` to clean up invalid entries
+
+### 7. Hooks Integration
+
+- Interactive setup for Claude Code hooks
+- pre_tool_use, post_tool_use, notification, stop hooks
+- Configurable via `setup-hooks` command
+
+## Configuration
+
+### Per-Project Role Override
+
+Create `.claude/roles.yaml`:
+
+```yaml
+roles:
+  tech-lead:
+    skills:
+      inherit: [dev-fe, dev-be, devops]  # Add devops to tech-lead
+
+  sa:
+    skills:
+      inherit: [dev-fe, dev-be]  # SA also knows code
+      include:
+        - category: architecture
+```
+
+## Research Sources
+
+Skills and architecture are built based on research from:
+
+1. **Agent-Skills-for-Context-Engineering** - Context engineering patterns
+2. **AI-Research-SKILLs** - 85 skills, 21 categories structure
+3. **everything-claude-code** - Hooks, commands, instinct system
+4. **BMAD-METHOD** - Agent YAML format, sharded workflows
+5. **spec-kit** - Extension system, multi-agent support
+
+## License
+
+MIT