Agentic System POC
A Proof of Concept agentic system built in Go demonstrating core agentic patterns: single-agent tool use and multi-agent orchestration.
Features
- Single Agent Mode: Interactive agent with calculator and file reader tools
Power your agent
การติดตั้งจากผู้เผยแพร่ ไม่ใช่เอนด์พอยต์ของ PlugMCP
URL แบบรีโมตหรือคำสั่งบนเครื่องของคุณ? การติดตั้งสองแบบต่างกันอย่างไร
รันบน PlugMCP
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
เปิดจากผู้เผยแพร่
A Proof of Concept agentic system built in Go demonstrating core agentic patterns: single-agent tool use and multi-agent orchestration.
A tour of building an ai agent with its own tool set, mcp client, and simple mcp server. Includes a single agent and multi agent example.
git clone <repository-url>
cd agentic-poc
go mod download
go build -o agent ./cmd/agent
Interactive mode with access to calculator and file reader tools:
./agent
# or explicitly:
./agent -mode single
Example interaction:
=== Single Agent Mode ===
Available tools: calculator, read_file
Type 'exit' or 'quit' to exit.
You: What is 15 + 27?
--- Intermediate Steps ---
[Tool Call] calculator
operation: add
a: 15
b: 27
Iterations: 2
---------------------------
Assistant: The sum of 15 and 27 is 42.
Architect creates a plan, Coder executes it:
./agent -mode multi
# With custom base path for file operations:
./agent -mode multi -path /tmp/workspace
Example interaction:
=== Multi-Agent Mode (Architect/Coder) ===
Enter a goal for the system to accomplish.
Type 'exit' or 'quit' to exit.
Goal: Create a hello world file
>>> Agent Transition: user -> architect
--- Plan ---
Goal: Create a hello world file
Steps:
1. Create hello.txt with greeting (action: write_file)
------------
>>> Agent Transition: architect -> coder
--- Actions Taken ---
• write_file: map[content:Hello, World! path:hello.txt]
---------------------
Summary: Successfully created hello.txt
Success: true
agentic-poc/
├── cmd/agent/ # CLI entry point
├── internal/
│ ├── provider/ # LLM provider abstraction
│ ├── tool/ # Tool interface and implementations
│ ├── memory/ # Conversation history
│ ├── agent/ # Agent loop and specialized agents
│ ├── orchestrator/ # Multi-agent coordination
│ └── cli/ # Command-line interface
├── test/integration/ # End-to-end tests
└── docs/wiki/ # Development learnings
For detailed C4 diagrams covering all operational modes, see docs/wiki/ARCHITECTURE.md.
User Input → Agent → LLM → Tool Call? → Execute Tool → LLM → Response
User Goal → Orchestrator → Architect Agent → Plan
→ Coder Agent → Execute Plan → Result
# All tests
go test ./...
# With verbose output
go test -v ./...
# Integration tests only
go test -v ./test/integration/
See docs/wiki/LEARNINGS.md for design decisions and challenges encountered during development.
See docs/wiki/OBSERVATIONS.md for high-level insights on building agentic systems - differentiation points, testing strategies, memory management, and observability.
MIT
A tour of building an ai agent with its own tool set, mcp client, and simple mcp server. Includes a single agent and multi agent example.
เครื่องมือ
ยังไม่มีรายการเครื่องมือ เซิร์ฟเวอร์แบบรีโมตจะได้รายการหลังการตรวจสอบสถานะครั้งแรก
FAQ
คัดลอกคอนฟิก stdio สำหรับ Cursor, Claude Desktop, Claude Code หรือ VS Code ได้ในหน้านี้ คอนฟิกนี้ใช้การติดตั้งของผู้เผยแพร่ ไม่ใช่พร็อกซีของ PlugMCP
ระเบียนนี้ไม่ได้ระบุตัวแปรสภาพแวดล้อมที่จำเป็น
Ai Agent Roadmap ทำงานแบบโลคัลผ่าน stdio
ตรวจสอบล่าสุดเมื่อ 2026-10-28 A tour of building an ai agent with its own tool set, mcp client, and simple mcp server. Includes a single agent and multi agent example.