@agentskillkit/agent-skills 1.0.1 → 3.2.1

package/.agent/workflows/architect.md

@@ -14,6 +14,98 @@ Create comprehensive project plans with task breakdown, architecture decisions,
 
 ---
 
+## 🛠️ Architecture Tools
+
+### 1. Diagram Generator
+
+Generate architecture diagrams with Mermaid:
+
+```mermaid
+# Context Diagram (C4 Level 1)
+graph TD
+    User[👤 User] --> App[📱 Application]
+    App --> DB[(🗄️ Database)]
+    App --> API[🌐 External API]
+```
+
+```mermaid
+# Container Diagram (C4 Level 2)
+graph TD
+    subgraph Frontend
+        Web[Next.js App]
+        Mobile[React Native]
+    end
+    subgraph Backend
+        API[API Gateway]
+        Auth[Auth Service]
+        Core[Core Service]
+    end
+    subgraph Data
+        DB[(PostgreSQL)]
+        Cache[(Redis)]
+    end
+    Web --> API
+    Mobile --> API
+    API --> Auth
+    API --> Core
+    Core --> DB
+    Core --> Cache
+```
+
+### 2. Dependency Analyzer
+
+Check project dependencies:
+```bash
+# NPM projects
+npx depcheck
+npx npm-check-updates
+
+# Visualize dependencies
+npx dependency-cruiser src --output-type dot | dot -T svg > deps.svg
+```
+
+### 3. Architecture Decision Records (ADR)
+
+Generate ADR template:
+```markdown
+# ADR-001: [Decision Title]
+
+## Status
+Proposed | Accepted | Deprecated | Superseded
+
+## Context
+What is the issue that we're seeing that is motivating this decision?
+
+## Decision
+What is the change that we're proposing and/or doing?
+
+## Consequences
+What becomes easier or more difficult because of this change?
+```
+
+---
+
+## 🤖 Meta-Agents Integration
+
+| Phase | Agent | Action |
+| ----- | ----- | ------ |
+| **Risk Evaluation** | `assessor` | Evaluate architecture risk before approval |
+| **Execution Planning** | `orchestrator` | Plan parallel vs sequential execution |
+| **Pattern Learning** | `learner` | Learn from past architecture decisions |
+
+```
+Flow:
+create plan → assessor.evaluate(plan)
+       ↓
+risk level? → add mitigations to plan
+       ↓
+orchestrator.plan(execution_order)
+       ↓
+handoff to /build with execution plan
+```
+
+---
+
 ## 🔴 MANDATORY: 4-Phase Planning
 
 ### Phase 1: Requirements Discovery
@@ -54,6 +146,37 @@ Level 4: Subtasks (atomic units)
 
 ---
 
+## 📐 Architecture Patterns Reference
+
+### Recommended Patterns
+
+| Pattern | When to Use | Complexity |
+|---------|-------------|------------|
+| **Monolith** | MVP, small team, fast iteration | Low |
+| **Microservices** | Large team, independent scaling | High |
+| **Serverless** | Event-driven, variable load | Medium |
+| **Modular Monolith** | Growing team, future split | Medium |
+
+### Data Patterns
+
+| Pattern | Use Case |
+|---------|----------|
+| **Repository** | Clean data access layer |
+| **CQRS** | Read/write optimization |
+| **Event Sourcing** | Audit trail, complex domain |
+| **Saga** | Distributed transactions |
+
+### Security Patterns
+
+| Pattern | Implementation |
+|---------|----------------|
+| **Authentication** | JWT, OAuth2, Session |
+| **Authorization** | RBAC, ABAC, PBAC |
+| **API Security** | Rate limiting, CORS, CSP |
+| **Data Protection** | Encryption at rest, TLS |
+
+---
+
 ## Output: PLAN.md
 
 Generated file: `docs/PLAN-{slug}.md`
@@ -126,41 +249,31 @@ After plan approval:
 
 ---
 
-## Naming Convention
-
-| Request | Generated File |
-|---------|----------------|
-| `/architect e-commerce app` | `docs/PLAN-ecommerce-app.md` |
-| `/architect user dashboard` | `docs/PLAN-user-dashboard.md` |
-| `/architect mobile API` | `docs/PLAN-mobile-api.md` |
-
-Rules:
-- Lowercase
-- Hyphen-separated
-- Max 30 characters
-- No special characters
-
----
-
-## Examples
-
-```
-/architect SaaS analytics dashboard
-/architect e-commerce with payments
-/architect mobile app backend
-/architect real-time chat system
-/architect portfolio website
-```
-
----
-
-## Key Principles
-
-1. **No code** - plan only, build later
-2. **User approval** - get sign-off before building
-3. **Realistic estimates** - include buffer time
-4. **Clear dependencies** - order tasks correctly
-5. **Verifiable** - each task has completion criteria
+## 🔧 Tech Stack Quick Reference
+
+### Frontend
+| Stack | Best For | Learning Curve |
+|-------|----------|----------------|
+| Next.js | Full-stack, SEO | Medium |
+| Vite + React | SPA, fast build | Low |
+| Remix | Forms, nested routes | Medium |
+| Astro | Content sites | Low |
+
+### Backend
+| Stack | Best For | Language |
+|-------|----------|----------|
+| Hono | Edge, lightweight | TypeScript |
+| Express | Flexibility, ecosystem | JavaScript |
+| FastAPI | ML, async, docs | Python |
+| NestJS | Enterprise, structure | TypeScript |
+
+### Database
+| Type | Options | Use Case |
+|------|---------|----------|
+| SQL | PostgreSQL, Supabase | Structured data |
+| NoSQL | MongoDB, Firebase | Flexible schema |
+| Graph | Neo4j | Relationships |
+| Vector | Pinecone, pgvector | AI/embeddings |
 
 ---
 
@@ -191,10 +304,3 @@ Next steps:
 | Need changes | Edit PLAN.md | Modify plan |
 | Complex project | `/autopilot` | Full orchestration |
 
-**Handoff to /build:**
-```markdown
-Plan approved. Run /build with context:
-- PLAN: docs/PLAN-{slug}.md
-- Stack: [selected stack]
-- Priority: [first epic]
-```

package/.agent/workflows/auto-accept-process.md

@@ -0,0 +1,275 @@
+---
+description: Autonomous agent execution framework. User approves PLAN.md once, agent runs end-to-end automatically.
+---
+
+# Auto-Accept Agent Workflow
+
+> **Mục tiêu:** User request → Agent chạy tự động → Kết quả hoàn chỉnh
+
+---
+
+## 🎯 Core Concept
+
+```
+User Request → PLAN.md (1 lần approve) → Agent chạy tự động → Done
+```
+
+**Thay vì:** User phải approve từng command, từng file edit.
+
+---
+
+## 🤖 Meta-Agents Integration
+
+| Phase | Agent | Action |
+| ----- | ----- | ------ |
+| **Pre-Execution** | `assessor` | Evaluate auto-execution risk level |
+| **Pre-Execution** | `recovery` | Save state before auto-run |
+| **During Execution** | `orchestrator` | Control parallel execution, retry |
+| **On Failure** | `recovery` | Auto-rollback to saved state |
+| **Post-Execution** | `learner` | Learn from execution patterns |
+
+```
+Flow:
+PLAN.md approved → assessor.evaluate(risk)
+       ↓
+recovery.save() → orchestrator.execute(phases)
+       ↓
+each phase → auto-accept if in allowPatterns
+       ↓
+failure? → recovery.restore() → learner.log()
+       ↓
+success → learner.log(patterns)
+```
+
+---
+
+## 📋 Quy Trình 3 Phase
+
+### Phase 1: Planning (1 Checkpoint)
+
+```markdown
+1. User: "Build feature X"
+2. Agent: Tạo PLAN.md với đầy đủ chi tiết
+3. ⛔ **CHECKPOINT DUY NHẤT**: User approve PLAN.md
+4. User: "Approved" hoặc "Proceed"
+```
+
+**PLAN.md phải có:**
+
+- [ ] Mục tiêu rõ ràng
+- [ ] Danh sách files sẽ tạo/sửa
+- [ ] Commands sẽ chạy
+- [ ] Verification steps
+
+### Phase 2: Execution (Auto-Accept)
+
+```markdown
+// turbo-all ← Magic annotation
+
+5. Agent tạo files (không hỏi)
+6. Agent chạy commands (auto-accept nếu safe)
+7. Agent fix lỗi nếu có
+```
+
+**Auto-Accept Conditions:**
+
+| Command Type          | Auto-Accept?      |
+| --------------------- | ----------------- |
+| `npm run *`           | ✅ Yes            |
+| `npm test`            | ✅ Yes            |
+| `git status/diff/log` | ✅ Yes            |
+| `node .agent/*`       | ✅ Yes            |
+| `git push`            | ❌ No (deny list) |
+| `rm -rf`              | ❌ Never          |
+
+### Phase 3: Verification (Auto)
+
+```markdown
+8. Agent chạy tests
+9. Agent chạy lint/security scan
+10. Agent report kết quả
+```
+
+---
+
+## 🔑 Magic Annotations
+
+### Trong Workflow Files
+
+```markdown
+// turbo ← Auto-accept STEP tiếp theo
+// turbo-all ← Auto-accept TẤT CẢ steps trong workflow
+```
+
+### Trong Code Blocks
+
+````markdown
+// @auto @safe ← Đánh dấu command an toàn
+
+```bash
+npm run test
+```
+````
+
+````
+
+---
+
+## ⚙️ Config: execution-policy.json
+
+```json
+{
+  "autoAccept": {
+    "enabled": true,
+    "defaultMode": "prompt",
+    "requiredAnnotations": ["auto", "safe"],
+
+    "allowPatterns": [
+      { "pattern": "^npm run [a-z0-9:-]+$", "type": "regex" },
+      { "pattern": "node .agent/", "type": "prefix" },
+      { "pattern": "git status", "type": "exact" }
+    ],
+
+    "denyPatterns": [
+      { "pattern": "rm -rf", "type": "contains", "severity": "critical" },
+      { "pattern": "git push", "type": "prefix", "severity": "high" }
+    ]
+  },
+
+  "phaseGate": {
+    "requirePlanApproval": true
+  }
+}
+````
+
+---
+
+## 📝 Template: User Request Format
+
+**Để trigger auto-accept mode:**
+
+```markdown
+/autopilot Build [feature description]
+
+Yêu cầu:
+
+1. [Chi tiết yêu cầu]
+2. [Constraints/preferences]
+
+AUTO-APPROVE: After PLAN.md approval, proceed without asking.
+```
+
+**Hoặc đơn giản:**
+
+```markdown
+/build [feature] --auto
+```
+
+---
+
+## 🔄 Flow Diagram
+
+```
+┌─────────────────┐
+│  User Request   │
+└────────┬────────┘
+         ▼
+┌─────────────────┐
+│   PLAN.md       │
+│   Generation    │
+└────────┬────────┘
+         ▼
+┌─────────────────┐
+│ ⛔ USER APPROVE │  ← Checkpoint duy nhất
+└────────┬────────┘
+         ▼
+┌─────────────────────────────────────┐
+│         AUTO-EXECUTION              │
+│  ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐   │
+│  │File │ │Cmd  │ │Test │ │Fix  │   │
+│  │Edit │→│Run  │→│Run  │→│Errs │   │
+│  └─────┘ └─────┘ └─────┘ └─────┘   │
+│         (All // turbo-all)          │
+└────────┬────────────────────────────┘
+         ▼
+┌─────────────────┐
+│    REPORT       │
+│    to User      │
+└─────────────────┘
+```
+
+---
+
+## ✅ Checklist để Enable Auto-Accept
+
+1. [ ] Workflow file có `// turbo-all` annotation
+2. [ ] Commands nằm trong `allowPatterns` của policy
+3. [ ] User đã approve PLAN.md
+4. [ ] Không có commands trong `denyPatterns`
+
+---
+
+## 🚀 Ví Dụ Thực Tế
+
+**User:**
+
+```
+/autopilot Build authentication system with JWT
+
+AUTO-APPROVE: After PLAN.md, proceed automatically.
+```
+
+**Agent Response:**
+
+```
+📋 PLAN.md created. Includes:
+- auth.service.ts (JWT logic)
+- auth.controller.ts (endpoints)
+- auth.middleware.ts (guards)
+- 12 test cases
+
+⛔ Approve to proceed?
+```
+
+**User:**
+
+```
+Proceed
+```
+
+**Agent (Auto from here):**
+
+```
+✅ Created auth.service.ts
+✅ Created auth.controller.ts
+✅ Created auth.middleware.ts
+✅ npm run test → 12/12 passed
+✅ npm run lint → No errors
+✅ Security scan → Clean
+
+🎉 Authentication system complete!
+```
+
+---
+
+## 📁 Files Liên Quan
+
+| File                                    | Purpose              |
+| --------------------------------------- | -------------------- |
+| `.agent/config/execution-policy.json`   | Allow/deny rules     |
+| `.agent/workflows/*.md`                 | Workflow definitions |
+| `.agent/scripts-js/execution-policy.js` | Policy engine        |
+| `.agent/scripts-js/workflow-engine.js`  | Workflow executor    |
+
+---
+
+## 🛡️ Safety Guarantees
+
+1. **Plan Approval Required**: Không chạy mà không có PLAN.md approved
+2. **Deny List Protected**: Dangerous commands luôn bị block
+3. **Execution History**: Mọi command được log
+4. **Rollback Ready**: Git commits trước mỗi phase
+
+---
+
+_Version 1.0 | Agent Skill Kit_