add docs.

Author: JDOxygen

.gitignore

@@ -244,4 +244,5 @@ CLAUDE.md
 # 运行时产生的文件
 cache_dir/
 local_file/
-deployment/
+deployment/
+docs_app/

docs/docs_en/api/agent/base_agent.md

@@ -24,7 +24,7 @@ The position of the class is:
 
 ---
 
-## Introduce
+## Introduction
 
 `BaseAgent` is the base class of all the agents. 
 

docs/docs_en/api/agent/base_flow.md

@@ -25,7 +25,7 @@ The position of the class is:
 
 ---
 
-## Introduce
+## Introduction
 
 `BaseFlow` is the base flow module for OxyGent framework.
 

docs/docs_en/api/agent/base_oxy.md

@@ -23,7 +23,7 @@ The position of the class is:
 ```
 ---
 
-## Introduce
+## Introduction
 `Oxy` is the abstract base class for all agents and tools in the OxyGent system.
 
 This class defines the core execution lifecycle, permission management,message handling, and data persistence patterns. It provides a unifiedinterface for both local and remote execution with comprehensive logging and error handling capabilities.

docs/docs_en/api/readme.md

@@ -1,92 +1,94 @@
-# OxyGent API Document
+# OxyGent API Reference
+
+> Detailed class documentation for all OxyGent components. Each page includes parameters, methods, inheritance hierarchy, and usage examples.
 
 ## Agents
 ---
-+ [BaseOxy](./agent/base_oxy.md)
-+ [BaseFlow](./agent/base_flow.md)
-+ [BaseAgent](./agent/base_agent.md)
-+ [LocalAgent](./agent/local_agent.md)
-+ [ChatAgent](./agent/chat_agent.md)
-    + [RAGAgent](./agent/rag_agent.md)
-+ [ReActAgent](./agent/react_agent.md)
-    + [ShellUseAgent](./agent/shell_use_agent.md)
-    + [SkillAgent](./agent/skill_agent.md)
-+ [ParallelAgent](./agent/parallel_agent.md)
-+ [WorkflowAgent](./agent/workflow_agent.md)
-+ [PlanAndSolveAgent](./agent/plan_and_solve_agent.md)
-+ [RemoteAgent](./agent/remote_agent.md)
-    + [SSEOxyGent](./agent/sse_oxy_agent.md)
-    + [A2AClientAgent](./agent/a2a_client_agent.md)
++ [BaseOxy](./agent/base_oxy.md) — The universal base class for all OxyGent components
++ [BaseFlow](./agent/base_flow.md) — Base class for execution flows
++ [BaseAgent](./agent/base_agent.md) — Base class for all agents
++ [LocalAgent](./agent/local_agent.md) — Agent that runs locally with an LLM
++ [ChatAgent](./agent/chat_agent.md) — Single-turn conversational agent
+    + [RAGAgent](./agent/rag_agent.md) — Retrieval-augmented generation agent
++ [ReActAgent](./agent/react_agent.md) — Reasoning + Acting loop with tool calling
+    + [ShellUseAgent](./agent/shell_use_agent.md) — SSH remote command execution
+    + [SkillAgent](./agent/skill_agent.md) — Dynamically loaded skill agent
++ [ParallelAgent](./agent/parallel_agent.md) — Parallel sub-task execution
++ [WorkflowAgent](./agent/workflow_agent.md) — Custom step-by-step workflows
++ [PlanAndSolveAgent](./agent/plan_and_solve_agent.md) — Plan first, then execute
++ [RemoteAgent](./agent/remote_agent.md) — Connect to agents in other processes
+    + [SSEOxyGent](./agent/sse_oxy_agent.md) — SSE-based distributed agent
+    + [A2AClientAgent](./agent/a2a_client_agent.md) — A2A protocol client agent
 
 ## Tools
 ---
-+ [BaseTool](./tools/base_tools.md)
-+ [HttpTool](./api_tools/http_tool.md)
-+ [FunctionTool](./function_tools/function_tool.md)
-+ [FunctionHub](./function_tools/function_hub.md)
-+ [MCPTool](./tools/mcp_tool.md)
-+ [BaseMCPClient](./tools/base_mcp_client.md)
-    + [StdioMCPClient](./tools/stdio_mcp_client.md) 
-    + [SSEMCPClient](./tools/sse_mcp_client.md)
-    + [StreamableMCPClient](./tools/streamable_mcp_client.md)
-+ [BaseBank](./bank_tools/base_bank.md)
-    + [BankTool](./bank_tools/bank_tool.md)
-    + [BankClient](./bank_tools/bank_client.md)
++ [BaseTool](./tools/base_tools.md) — Base class for all tools
++ [HttpTool](./api_tools/http_tool.md) — HTTP API wrapper tool
++ [FunctionTool](./function_tools/function_tool.md) — Single Python function as a tool
++ [FunctionHub](./function_tools/function_hub.md) — Collection of Python function tools
++ [MCPTool](./tools/mcp_tool.md) — MCP protocol tool wrapper
++ [BaseMCPClient](./tools/base_mcp_client.md) — Base MCP client
+    + [StdioMCPClient](./tools/stdio_mcp_client.md) — MCP over stdio transport
+    + [SSEMCPClient](./tools/sse_mcp_client.md) — MCP over SSE transport
+    + [StreamableMCPClient](./tools/streamable_mcp_client.md) — MCP over Streamable HTTP
++ [BaseBank](./bank_tools/base_bank.md) — Base class for tool banks
+    + [BankTool](./bank_tools/bank_tool.md) — Server-side tool bank
+    + [BankClient](./bank_tools/bank_client.md) — Client for remote tool banks
 
 ## Preset Tools
 ---
-+ [Preset Tools (10 FunctionHub collections)](./preset_tools.md)
++ [Preset Tools (10 FunctionHub collections)](./preset_tools.md) — Built-in tools: file, math, time, shell, python, http, image_gen, ssh, string, system
 
-## Flow
+## Flows
 ---
-+ [WorkFlow](./flows/workflow.md)
-+ [ParallelFlow](./flows/parallel_flow.md)
-+ [PlanAndSolve](./flows/plan_and_solve.md)
-+ [Reflexion (+ MathReflexion)](./flows/reflexion.md)
++ [WorkFlow](./flows/workflow.md) — Sequential step execution
++ [ParallelFlow](./flows/parallel_flow.md) — Concurrent step execution
++ [PlanAndSolve](./flows/plan_and_solve.md) — Decompose → plan → solve
++ [Reflexion (+ MathReflexion)](./flows/reflexion.md) — Self-reflection and retry loop
 
 ## LLM
 ---
-+ [BaseLLM](./llms/base_llm.md)
-+ [RemoteLLM](./llms/remote_llm.md)
-    + [HttpLLM](./llms/http_llm.md)
-    + [OpenAILLM](./llms/openai_llm.md)
-+ [LocalLLM](./llms/local_llm.md)
-+ [MockLLM](./llms/mock_llm.md)
++ [BaseLLM](./llms/base_llm.md) — Base class for all LLM integrations
++ [RemoteLLM](./llms/remote_llm.md) — Base for remote API-based LLMs
+    + [HttpLLM](./llms/http_llm.md) — Call models via HTTP API
+    + [OpenAILLM](./llms/openai_llm.md) — Call models via OpenAI SDK
++ [LocalLLM](./llms/local_llm.md) — Load HuggingFace models locally
++ [MockLLM](./llms/mock_llm.md) — Mock model for testing
 
 ## Database
 ---
-+ [BaseDB](./databases/base_db.md)
-+ [BaseES](./databases/db_es/base_es.md)
-    + [JesES](./databases/db_es/jes_es.md)
-    + [LocalES](./databases/db_es/local_es.md)
-    + [MemoryEs](./databases/db_es/memory_es.md)
-+ [BaseRedis](./databases/db_redis/base_redis.md)
-    + [JimdbApRedis](./databases/db_redis/jimdb_ap_redis.md)
-    + [LocalRedis](./databases/db_redis/local_redis.md)
-+ [BaseVectorDB](./databases/db_vector/base_vector_db.md)
-    + [VearchDB](./databases/db_vector/vearch_db.md)
++ [BaseDB](./databases/base_db.md) — Base database interface
++ [BaseES](./databases/db_es/base_es.md) — Elasticsearch base
+    + [JesES](./databases/db_es/jes_es.md) — JD Elasticsearch
+    + [LocalES](./databases/db_es/local_es.md) — File-based local ES fallback
+    + [MemoryEs](./databases/db_es/memory_es.md) — In-memory ES for testing
++ [BaseRedis](./databases/db_redis/base_redis.md) — Redis base
+    + [JimdbApRedis](./databases/db_redis/jimdb_ap_redis.md) — JD Redis
+    + [LocalRedis](./databases/db_redis/local_redis.md) — File-based local Redis fallback
++ [BaseVectorDB](./databases/db_vector/base_vector_db.md) — Vector database base
+    + [VearchDB](./databases/db_vector/vearch_db.md) — Vearch vector DB
 
 ## Schemas
 ---
-+ [OxyRequest / OxyResponse / OxyState](./schemas/oxy.md)
-+ [LLMState / LLMResponse](./schemas/llm.md)
-+ [Message / Memory](./schemas/memory.md)
-+ [Observation / ExecResult](./schemas/observation.md)
-+ [SkillMetadata](./schemas/skill.md)
-+ [Evaluation (Rating models)](./schemas/evaluation.md)
++ [OxyRequest / OxyResponse / OxyState](./schemas/oxy.md) — Core request/response types
++ [LLMState / LLMResponse](./schemas/llm.md) — LLM interaction types
++ [Message / Memory](./schemas/memory.md) — Conversation history and memory
++ [Observation / ExecResult](./schemas/observation.md) — Tool execution results
++ [SkillMetadata](./schemas/skill.md) — Skill definition metadata
++ [Evaluation (Rating models)](./schemas/evaluation.md) — Rating and evaluation types
 
 ## MAS System Modules
 ---
-+ [Config](./config.md)
-+ [MAS](./mas.md)
-+ [DBFactory](./db_factory.md)
-+ [EmbeddingCache](./embedding_cache.md)
-+ [OxyFactory](./oxy_factory.md)
++ [Config](./config.md) — Global configuration singleton
++ [MAS](./mas.md) — Multi-Agent System runtime container
++ [DBFactory](./db_factory.md) — Database connection factory
++ [EmbeddingCache](./embedding_cache.md) — Embedding vector cache
++ [OxyFactory](./oxy_factory.md) — Component factory for deserialization
 
 ## Live Prompt
 ---
-+ [Live Prompt (PromptManager, PromptOptimizer, VersionSync, DynamicAgentManager)](./live_prompt.md)
++ [Live Prompt](./live_prompt.md) — PromptManager, PromptOptimizer, VersionSync, DynamicAgentManager
 
 ## Transport
 ---
-+ [A2A Protocol (A2AServerGateway, A2AInMemoryStore)](./transport_a2a.md)
++ [A2A Protocol](./transport_a2a.md) — A2AServerGateway, A2AInMemoryStore

docs/docs_en/examples/readme.md

@@ -4,6 +4,18 @@ This directory contains documentation for all OxyGent example scripts, organized
 
 ---
 
+## Start Here
+
+If you're new to OxyGent, try these three examples first:
+
+| Example | What You'll Learn |
+|---------|------------------|
+| [Single Agent](./agents/demo_single_agent.md) | Minimal ChatAgent with custom prompts and hooks |
+| [ReAct Agent with Tools](./agents/demo_react_agent.md) | ReActAgent with reflexion and tool calling |
+| [Hierarchical Multi-Agent](./agents/demo_hierarchical_agents.md) | Master-sub agent architecture and delegation |
+
+---
+
 ## Categories
 
 | Category | Examples | Description |

docs/docs_en/introduction/a2a/demo-guide.md

@@ -104,3 +104,8 @@ Capability Descriptions:
 ## Related Examples
 
 - [A2A Examples Overview](../../examples/a2a/readme.md) -- Complete list of A2A-related examples with instructions
+
+---
+
+[Next: A2A Design & Capabilities](./design.md)
+[Back to Home](../readme.md)

docs/docs_en/introduction/a2a/design.md

@@ -130,3 +130,8 @@ For existing OxyGent users, A2A integration can be quickly deployed through "min
 ## Related Examples
 
 - [A2A Server Example](../../examples/a2a/demo_a2a_oxygent_server.md) -- Demonstrates how to start an OxyGent A2A Server
+
+---
+
+[Previous: A2A Quick Start](./demo-guide.md)
+[Back to Home](../readme.md)

docs/docs_en/introduction/advanced/multimodal.md

@@ -125,6 +125,7 @@ if __name__ == "__main__":
 
 
 [Previous: Retrieving Memory and Regenerating](./continue-exec.md)
+[Next: RAG](./rag.md)
 [Back to Home](../readme.md)
 
 ---

docs/docs_en/introduction/advanced/rag.md

@@ -1,4 +1,4 @@
-## How to Perform Retrieval-Augmented Generation?
+# How to Perform Retrieval-Augmented Generation?
 
 OxyGent supports injecting knowledge into prompts via the `knowledge` parameter. Below is a simple RAG example: