@nockdev/awf 6.2.0 → 6.2.1

package/.agent/rules/modules/terminal-safety.yaml

@@ -3,7 +3,7 @@
 
 ---
 name: terminal-safety
-version: "6.1.2"
+version: "6.2.1"
 rule_id: "MOD-TRM-001"
 
 # =============================================================================

package/.agent/rules/modules/yagni.yaml

@@ -3,7 +3,7 @@
 
 ---
 name: yagni-enforcement
-version: "6.1.2"
+version: "6.2.1"
 rule_id: "MOD-YAG-001"
 
 # =============================================================================

package/.agent/rules/validation-framework.md

@@ -6,7 +6,7 @@ category: quality
 version: "4.5"
 ---
 
-# ✅ Validation Framework v6.2.0
+# ✅ Validation Framework v6.2.1
 
 > 🌍 **Language / Ngôn ngữ**: English (default) | [Tiếng Việt](#tiếng-việt)
 > 📝 **Note**: Consolidates `pre-check-validation.md` + `architecture-aware-implementation.md`

package/.agent/skills/DEVELOPMENT.yaml

@@ -1,6 +1,6 @@
 # DOMYH Awesome Code — Skill Development Guide
 # How to create and maintain skills
-# Version: 6.1.2
+# Version: 6.2.1
 
 # =============================================================================
 # 📋 OVERVIEW
@@ -63,7 +63,7 @@ workflow:
       template: |
         name: new-skill
         display: New Skill Name
-        category: language  # or framework, devops, support
+        category: language  # or framework, devops, cross-cutting
         priority: 2
         desc: "Brief description under 80 chars"
         keywords:
@@ -150,10 +150,10 @@ categories:
     description: "Infrastructure and deployment"
     examples: [docker, kubernetes, aws, ci-cd]
 
-  support:
-    priority: 2
+  cross-cutting:
+    priority: 4
     description: "Cross-cutting concerns"
-    examples: [database, testing]
+    examples: [database, testing, tailwind, audit-pro]
 
 # =============================================================================
 # 🎯 BEST PRACTICES

package/.agent/skills/README.md

@@ -6,24 +6,27 @@
 
 ## Quick Reference
 
-| Metric          | Value                  |
-| --------------- | ---------------------- |
-| Total Skills    | 60                     |
-| Total META Size | 18.2KB                 |
-| Avg per skill   | 325 bytes (~81 tokens) |
-| Max active      | 3 skills               |
+| Metric          | Value                 |
+| --------------- | --------------------- |
+| Total Skills    | 82                    |
+| Total META Size | ~17KB                 |
+| Avg per skill   | ~300 bytes (~100 tok) |
+| Max detected    | 5 META.yaml           |
+| Max active T2   | 3 SKILL.md            |
 
 ---
 
 ## Categories
 
-| Category               | Skills                                                                                                     | Priority   |
-| ---------------------- | ---------------------------------------------------------------------------------------------------------- | ---------- |
-| **Core** (6)           | security, api-design, error-handling, logging, observability, authentication                               | 0 (always) |
-| **Language** (28)      | go, python, typescript, javascript, rust, cpp, c, csharp, java, php, kotlin, swift, lua, asm               | 1          |
-| **Framework** (8)      | react, vue, nextjs, nuxt, angular, svelte, flutter, react-native                                           | 2          |
-| **DevOps** (4)         | docker, kubernetes, aws, ci-cd                                                                             | 3          |
-| **Cross-cutting** (14) | database, testing, tailwind, ui-ux-pro-max, web-perf, sql, bun, deno, coding-rules, electron, react-native | 2          |
+| Category               | Skills                                                                                                | Priority   |
+| ---------------------- | ----------------------------------------------------------------------------------------------------- | ---------- |
+| **Core** (6)           | security, api-design, error-handling, logging, observability, authentication                          | 0 (always) |
+| **Language** (28)      | go, python, typescript, javascript, rust, cpp, c, csharp, java, php, kotlin, swift, lua, asm, ...     | 1          |
+| **Framework** (8)      | react, vue, nextjs, nuxt, angular, svelte, flutter, react-native                                      | 2          |
+| **DevOps** (7)         | docker, kubernetes, aws, ci-cd, terraform, gcp, azure                                                 | 3          |
+| **Cross-cutting** (21) | testing, database, sql, tailwind, electron, coding-rules, domyh-design, web-perf, playwright, ...     | 4          |
+| **Tooling** (5)        | mcp, api-protocols, ide-extension, cli-dev, browser-agent                                             | 5          |
+| **AI-ML** (7)          | ai-agents, prompt-engineering, rag-patterns, vector-search, gemini-media-gen, gemini-tts, gemini-live | 6          |
 
 ---
 
@@ -37,13 +40,13 @@
 
 ---
 
-## META.yaml Schema (v6.2.0)
+## META.yaml Schema (v6.2.1)
 
 ```yaml
 name: skill-id          # lowercase, hyphen-separated
 display: Display Name   # Human-readable
-category: language      # core|language|framework|devops|support
-priority: 1             # 0-5 (lower = higher)
+category: language      # core|language|framework|devops|cross-cutting|tooling|ai-ml
+priority: 1             # 0-6 (lower = higher)
 desc: "Short desc"      # Max 80 chars
 keywords: [kw1, kw2]    # Max 5-7 keywords
 detect: [*.ext, file]   # File patterns

package/.agent/skills/_categories.yaml

@@ -79,7 +79,7 @@ categories:
       - react-native
 
   # ==========================================================================
-  # DEVOPS (4 skills)
+  # DEVOPS (7 skills)
   # ==========================================================================
   devops:
     description: "Infrastructure and deployment"
@@ -90,6 +90,10 @@ categories:
       - kubernetes
       - aws
       - ci-cd
+      # v6.3 New
+      - terraform
+      - gcp
+      - azure
 
   # ==========================================================================
   # CORE (6 skills - always loaded)
@@ -108,7 +112,7 @@ categories:
       - authentication
 
   # ==========================================================================
-  # CROSS-CUTTING (10 skills)
+  # CROSS-CUTTING (21 skills)
   # ==========================================================================
   cross-cutting:
     description: "Cross-cutting concerns and utilities"
@@ -121,25 +125,73 @@ categories:
       - tailwind
       - electron
       - coding-rules
-      - ui-ux-pro-max
+      - domyh-design
       - web-perf
       - deno
       - bun
       - audit-pro
 
+      # P0 Skills (v6.2.1)
+      - tdd-workflow
+      - accessibility
+      - seo
+
+      # P1 Skills (v6.2.1)
+      - microservices
+      - monorepo
+
+      # v6.3 New
+      - tauri
+      - real-time
+      - event-driven
+      - wasm
+      - playwright
+
+  # ==========================================================================
+  # TOOLING (5 skills) — NEW v6.3
+  # ==========================================================================
+  tooling:
+    description: "Development tools, protocols, and extensions"
+    path: "tooling/"
+    priority: 2
+    skills:
+      - mcp
+      - ide-extension
+      - cli-dev
+      - api-protocols
+      - browser-agent
+
+  # ==========================================================================
+  # AI-ML (7 skills) — v6.3 + Google AI
+  # ==========================================================================
+  ai-ml:
+    description: "AI/ML agent systems, vector search, prompt engineering, and Gemini APIs"
+    path: "ai-ml/"
+    priority: 3
+    skills:
+      - ai-agents
+      - vector-search
+      # Migrated from cross-cutting
+      - prompt-engineering
+      - rag-patterns
+      # v6.3 New — Google AI
+      - gemini-media-gen
+      - gemini-tts
+      - gemini-live
+
 # =============================================================================
 # STATISTICS
 # =============================================================================
 
 stats:
-  total_skills: 56
-  categories: 5
+  total_skills: 82
+  categories: 7
   estimated_tokens:
-    all_meta: 5600 # ~100 tokens × 56
+    all_meta: 8200 # ~100 tokens × 82
     router_only: 500 # Router + core META
-    typical_session: 2500 # Router + 2 language + 1 framework
+    typical_session: 3000 # Router + 2 language + 1 framework + 1 tooling
 
 
 # =============================================================================
-# _DOMYH Awesome Code • Category Definitions_
+# _DOMYH Awesome Code • Category Definitions v6.3_
 # =============================================================================

package/.agent/skills/_router.yaml

@@ -54,20 +54,14 @@ strategy:
     - "ADVANCED.md" # Reference only when explicitly needed
 
 # =============================================================================
-# P0-P2 INTEGRATION (v1.1)
+# SELECTION CONFIG
 # =============================================================================
-# References core/ROUTER.yaml for full implementation
 
 selection:
-  # P0: Dynamic threshold (see core/ROUTER.yaml for full config)
-  threshold_config: "$ref:core/ROUTER.yaml#/selection/threshold"
-  default_threshold: 0.50
-
-  # P1: Query expansion (see core/ROUTER.yaml for synonyms)
-  query_expansion: "$ref:core/ROUTER.yaml#/query_expansion"
-
-  # P2: Confidence scoring
-  confidence_scoring: "$ref:core/ROUTER.yaml#/confidence_scoring"
+  default_threshold: 0.30
+  recency_boost: 0.15
+  max_active_skills: 5 # Max T1 META.yaml detected
+  max_active_t2: 3 # Max SKILL.md loaded simultaneously
 
 # =============================================================================
 # CATEGORY ROUTING
@@ -93,10 +87,10 @@ routing:
   devops:
     path: "skills/devops/"
     detect_by:
-      - "special_file" # Dockerfile, k8s/
+      - "special_file" # Dockerfile, k8s/, *.tf
       - "directory" # .github/workflows/
     priority: 3
-    count: 4
+    count: 7
 
   core:
     path: "skills/core/"
@@ -104,17 +98,27 @@ routing:
     priority: 0
     count: 6
 
+  private:
+    path: "private/"
+    detect_by:
+      - "keyword"
+      - "dependency"
+    priority: -1 # Highest priority — overrides public skills
+    count: 0 # Dynamic: read from private/_index.yaml
+    registry: "private/_index.yaml"
+    note: "Gitignored private skills, not in public repo"
+
   cross-cutting:
     path: "skills/cross-cutting/"
     detect_by:
       - "dependency"
       - "extension"
     priority: 4
-    count: 10
+    count: 21
 
-    # UI/UX Pro Max skill - integrated with HSA adapter
+    # HSA adapter overrides for cross-cutting skills
     skills:
-      ui-ux-pro-max:
+      domyh-design:
         adapter: "hsa.adapters.uiux_adapter"
         preload_data: true
         triggers:
@@ -138,7 +142,6 @@ routing:
             - "typography"
             - "font pairing"
 
-      # Electron skill - integrated with dedicated HSA YAML-based adapter
       electron:
         adapter: "hsa.adapters.electron_adapter"
         data_dir: "skills/cross-cutting/electron/data"
@@ -165,7 +168,6 @@ routing:
             - "better-sqlite3"
             - "electron-store"
 
-      # Tailwind skill - integrated with dedicated HSA YAML-based adapter
       tailwind:
         adapter: "hsa.adapters.tailwind_adapter"
         data_dir: "skills/cross-cutting/tailwind/data"
@@ -197,6 +199,22 @@ routing:
             - "lightning css"
             - "Vite plugin"
 
+  tooling:
+    path: "skills/tooling/"
+    detect_by:
+      - "special_file" # mcp.json, .vscode/extensions.json
+      - "dependency" # @modelcontextprotocol/sdk
+    priority: 5
+    count: 5
+
+  ai-ml:
+    path: "skills/ai-ml/"
+    detect_by:
+      - "dependency" # langchain, openai, pinecone
+      - "keyword" # RAG, embeddings, vector, gemini
+    priority: 6
+    count: 7
+
 # =============================================================================
 # SKILL RESOLUTION
 # =============================================================================
@@ -215,6 +233,29 @@ resolution:
     docker: "devops/docker"
     security: "core/security"
     testing: "cross-cutting/testing"
+    # v6.3 new aliases
+    mcp: "tooling/mcp"
+    graphql: "tooling/api-protocols"
+    trpc: "tooling/api-protocols"
+    grpc: "tooling/api-protocols"
+    ai-agents: "ai-ml/ai-agents"
+    langchain: "ai-ml/ai-agents"
+    vector-search: "ai-ml/vector-search"
+    tauri: "cross-cutting/tauri"
+    websocket: "cross-cutting/real-time"
+    sse: "cross-cutting/real-time"
+    kafka: "cross-cutting/event-driven"
+    rabbitmq: "cross-cutting/event-driven"
+    nats: "cross-cutting/event-driven"
+    terraform: "devops/terraform"
+    gcp: "devops/gcp"
+    azure: "devops/azure"
+    wasm: "cross-cutting/wasm"
+    webassembly: "cross-cutting/wasm"
+    playwright: "cross-cutting/playwright"
+    e2e: "cross-cutting/playwright"
+    browser-testing: "cross-cutting/playwright"
+    browser-agent: "tooling/browser-agent"
 
 # =============================================================================
 # TOKEN BUDGET
@@ -228,5 +269,6 @@ token_budget:
 
 
 # =============================================================================
-# _DOMYH Awesome Code • Hierarchical Skill Router_
+# _DOMYH Awesome Code • Hierarchical Skill Router v6.3_
 # =============================================================================
+

package/.agent/skills/ai-ml/ai-agents/META.yaml

@@ -0,0 +1,127 @@
+name: ai-agents
+version: "1.1.0"
+display: AI Agent Orchestration
+category: ai-ml
+tier: 1
+priority: 1
+
+desc: "AI agent development & orchestration. LangChain/LangGraph, CrewAI, AutoGen, Google ADK. Covers multi-agent patterns, LLM API integration (OpenAI/Anthropic/Gemini), MCP built-in support, Gemini Live agents, tool use, memory, and production deployment."
+
+  ### Data Files (4 domains, ~32 patterns)
+  - orchestration-patterns.yaml: supervisor/swarm/sequential/hierarchical (8)
+  - llm-integration.yaml: OpenAI/Anthropic/Gemini API patterns, streaming (8)
+  - agent-rules.yaml: Cursor/Cline/Codex/DOMYH rule system comparison (8)
+  - memory-patterns.yaml: short-term, long-term, vector memory (8)
+
+triggers:
+  file_patterns:
+    # LangChain/LangGraph
+    - "**/agents/**"
+    - "**/chains/**"
+    - "langgraph.json"
+    # Config
+    - "AGENTS.md"
+    - ".clinerules"
+    - ".cursor/rules/**"
+  keywords:
+    # Frameworks
+    - ai agent
+    - langchain
+    - langgraph
+    - crewai
+    - autogen
+    - microsoft autogen
+    - agent framework
+    # Patterns
+    - multi-agent
+    - agent orchestration
+    - supervisor pattern
+    - swarm pattern
+    - sequential chain
+    - hierarchical agents
+    # LLM Integration
+    - tool use
+    - function calling
+    - structured output
+    - LLM integration
+    - OpenAI API
+    - Anthropic API
+    - Gemini API
+    - Claude API
+    - streaming response
+    - token optimization
+    - prompt engineering
+    # Memory
+    - agent memory
+    - conversation memory
+    - vector memory
+    - long-term memory
+    # Agent rules
+    - AGENTS.md
+    - cursor rules
+    - cline rules
+    - agent skills
+    # Observability
+    - langsmith
+    - langfuse
+    - agent traces
+    # Production
+    - human-in-the-loop
+    - HITL
+    # v1.1 — Gemini ecosystem
+    - google adk
+    - agent development kit
+    - gemini live agent
+    - MCP built-in
+    - a2a protocol
+
+caps:
+  # Frameworks
+  - LangChain Expression Language (LCEL) chains
+  - LangGraph stateful workflows (nodes/edges/state)
+  - CrewAI role-based multi-agent collaboration
+  - AutoGen conversational agents
+  # Orchestration
+  - Supervisor pattern (single orchestrator)
+  - Swarm pattern (peer-to-peer agents)
+  - Sequential pipeline (chain of agents)
+  - Hierarchical delegation (manager → workers)
+  # LLM Integration
+  - Multi-provider LLM integration (OpenAI, Anthropic, Gemini, local)
+  - Function calling / Tool use across providers
+  - Structured output (JSON mode, Pydantic models)
+  - Streaming responses (SSE, async generators)
+  - Token optimization & context window management
+  # Memory
+  - Short-term memory (conversation buffer)
+  - Long-term memory (vector store backed)
+  - Episodic memory (event-based recall)
+  # Agent Configuration
+  - Cursor Rules system (global → project → dynamic .mdc)
+  - Cline Rules (.clinerules/ folder)
+  - Codex AGENTS.md configuration
+  - DOMYH skill system (META.yaml → SKILL.md → data/)
+  # Production
+  - Human-in-the-loop (HITL) approval patterns
+  - Observability (LangSmith, Langfuse, traces/spans)
+  - Async queue architecture for long-running agents
+  - Rate limiting & cost optimization
+  - Error recovery & retry strategies
+  # v1.1
+  - Google ADK (Agent Development Kit)
+  - Gemini Live real-time agents (Pipecat, LiveKit, ADK)
+  - MCP built-in tool integration
+  - A2A (Agent-to-Agent) protocol
+
+related_skills:
+  - mcp
+  - prompt-engineering
+  - rag-patterns
+  - vector-search
+  - gemini-live
+  - gemini-media-gen
+
+hsa_adapter:
+  enabled: true
+  data_dir: "skills/ai-ml/ai-agents/data"
+  preload_data: true

package/.agent/skills/ai-ml/ai-agents/SKILL.md

@@ -0,0 +1,139 @@
+# AI Agent Orchestration — LangChain • CrewAI • AutoGen
+
+> Multi-agent systems • LLM API integration • Agent rules  
+> Tool use • Memory management • Production patterns
+
+---
+
+## Khi Nào Dùng
+
+- Xây dựng AI agents (single/multi-agent)
+- Integrate LLM APIs (OpenAI, Anthropic, Gemini)
+- Thiết kế agent orchestration patterns
+- Implement agent memory & tool use
+- Cấu hình agent rules (Cursor, Cline, AGENTS.md)
+
+## Orchestration Patterns
+
+```
+┌─────────────────────────────────────────────────────┐
+│ SUPERVISOR (recommended for complex tasks)          │
+│                                                     │
+│  ┌──────────┐                                       │
+│  │Supervisor│──┬──▶ Agent A (Research)               │
+│  │(Router)  │  ├──▶ Agent B (Code)                   │
+│  └──────────┘  └──▶ Agent C (Review)                 │
+│                                                     │
+│ SWARM (for collaborative tasks)                     │
+│  Agent A ◄──▶ Agent B ◄──▶ Agent C                   │
+│                                                     │
+│ SEQUENTIAL (for pipeline tasks)                     │
+│  Agent A ──▶ Agent B ──▶ Agent C                     │
+└─────────────────────────────────────────────────────┘
+```
+
+| Pattern          | Use Case                   | Complexity         |
+| ---------------- | -------------------------- | ------------------ |
+| **Sequential**   | ETL, simple pipelines      | ⭐ Low             |
+| **Supervisor**   | Complex tasks, routing     | ⭐⭐ Medium        |
+| **Swarm**        | Collaborative, peer review | ⭐⭐⭐ High        |
+| **Hierarchical** | Enterprise, delegation     | ⭐⭐⭐⭐ Very High |
+
+## LLM API Integration
+
+### Multi-Provider Pattern
+
+```typescript
+// Provider-agnostic interface
+interface LLMProvider {
+  chat(messages: Message[], options?: Options): Promise<Response>;
+  stream(messages: Message[], options?: Options): AsyncGenerator<Chunk>;
+}
+
+// Structured output (all providers)
+const result = await llm.chat(messages, {
+  responseFormat: { type: "json_schema", schema: userSchema },
+});
+```
+
+### Function Calling
+
+```typescript
+// Tool definition (OpenAI format, widely adopted)
+const tools = [
+  {
+    type: "function",
+    function: {
+      name: "search_database",
+      description: "Search the product database",
+      parameters: {
+        type: "object",
+        properties: {
+          query: { type: "string" },
+          limit: { type: "number", default: 10 },
+        },
+        required: ["query"],
+      },
+    },
+  },
+];
+```
+
+### Streaming
+
+```typescript
+// SSE streaming pattern
+for await (const chunk of llm.stream(messages)) {
+  process.stdout.write(chunk.content); // real-time output
+}
+```
+
+## Agent Rules Systems
+
+| System     | Location              | Format                       | Scope            |
+| ---------- | --------------------- | ---------------------------- | ---------------- |
+| **Cursor** | `.cursor/rules/*.mdc` | MDC (frontmatter + markdown) | Dynamic per-file |
+| **Cursor** | Settings → Rules      | Plain text                   | Global user      |
+| **Cline**  | `.clinerules/`        | Markdown files               | Project          |
+| **Codex**  | `AGENTS.md`           | Markdown                     | Project          |
+| **DOMYH**  | `.agent/skills/`      | YAML + Markdown              | Hierarchical     |
+
+### Best Practices
+
+1. **Concise** — Under 500 lines per rule file
+2. **Specific** — Show patterns, not just principles
+3. **Scoped** — Use glob patterns for context-specific rules
+4. **Layered** — Global → Project → File-specific
+
+## Memory Patterns
+
+| Type         | Duration   | Storage     | Use Case              |
+| ------------ | ---------- | ----------- | --------------------- |
+| **Buffer**   | Session    | In-memory   | Chat history          |
+| **Summary**  | Session    | In-memory   | Long conversations    |
+| **Vector**   | Persistent | Vector DB   | Recall past knowledge |
+| **Episodic** | Persistent | DB + Vector | Event-based recall    |
+
+## Production Checklist
+
+- ✅ Rate limiting per provider API key
+- ✅ Retry with exponential backoff
+- ✅ Token usage tracking & cost alerts
+- ✅ Async queue for long-running agents
+- ✅ Human-in-the-loop for destructive actions
+- ✅ Observability: traces, spans, token counts
+- ✅ Fallback provider on rate limit
+
+## Common Traps
+
+| Trap                    | Fix                                            |
+| ----------------------- | ---------------------------------------------- |
+| Context window overflow | Summarize history, trim old messages           |
+| Agent loops             | Max iterations limit, break conditions         |
+| Hallucination in tools  | Validate tool outputs, constrained prompts     |
+| High costs              | Cache responses, use smaller models for triage |
+| Prompt injection        | Input sanitization, system prompt protection   |
+
+---
+
+_DOMYH Awesome Code • AI Agents Skill v1.0.0_