A comprehensive collection of production-ready AI agent implementations showcasing different frameworks, protocols, and integration patterns. This repository demonstrates various approaches to building intelligent agents with Model Context Protocol (MCP), multi-agent systems, and real-world integrations.
Repository Overview
Analyze stocks with summaries, price targets, and analyst recommendations. Track SEC filings, divi…
2026-11-13
อัปไทม์การตรวจสอบ 1 ครั้งล่าสุด
ดาว
—
เครื่องมือ
—
อัปไทม์
0%
This repository contains multiple agent implementations, each demonstrating different architectural patterns and use cases:
Timeouts: Extended timeout handling (ADK) for long operations
Specialization: SEO research, e-commerce intelligence, market analysis
Technical Stack:
LangGraph Agent: OpenAI GPT with MemorySaver checkpointer
ADK Agent: Google Gemini 2.0 Flash with custom timeout patches
MCP Integration: Bright Data MCP server for data collection
Slack Integration: Bot with agent selection and interactive UI
Agent Comparison:
Feature
LangGraph Agent
ADK Agent
Memory
Persistent (checkpointer)
Context-aware (5 messages)
Timeout
Standard (5s)
Extended (60s)
Model
OpenAI GPT
Gemini 2.0 Flash
Best For
Interactive conversations
Long-running operations
Use Cases:
SEO keyword research and SERP analysis
E-commerce product monitoring and price tracking
Competitor analysis and market intelligence
Web scraping and data collection
Business intelligence and insights
fpl-deepagent/
Fantasy Premier League MCP Assistant
A comprehensive Fantasy Premier League assistant that integrates with ChatGPT through the Model Context Protocol (MCP), featuring beautiful React UI components and real-time FPL data.
Key Features:
Framework: FastMCP with Streamable HTTP transport
UI Integration: React 18 + TypeScript components for ChatGPT
Real-time Data: Live FPL API integration with caching and error handling
A production-ready tutorial showing how to build a ChatGPT App with:
FastMCP (Streamable HTTP) as the MCP server
React widgets rendered inside ChatGPT
Supabase (Postgres) as authoritative state for tasks/notifications
OAuth (Auth0) for multi-user authentication (MCP OAuth)
Optional Slack notifications (send now + schedule)
Start here:
task-manager-app/README.md
notion-mcp-agent/
Notion Knowledge Management Agent
A sophisticated agent that integrates with Notion through MCP, providing intelligent database management and knowledge organization capabilities.
Key Features:
Framework: LangGraph with MCP integration
Integration: Notion API for database operations
Slack Interface: Interactive knowledge management
Context Management: Intelligent data aggregation
Database Operations: Create, read, update, and organize Notion databases
Technical Stack:
LangGraph for agent orchestration
Notion MCP server for database operations
Slack API for user interaction
Context aggregation for intelligent responses
Use Cases:
Knowledge base management
Database organization and maintenance
Content aggregation and structuring
Team collaboration workflows
Information retrieval and organization
claude-advanced-tool-use/
Claude Advanced Tool Use Tutorial
A comprehensive tutorial demonstrating Anthropic's Advanced Tool Use features: Programmatic Tool Calling (PTC) and Tool Search. These features enable AI agents to scale to thousands of tools while dramatically reducing token usage.
Key Features:
Programmatic Tool Calling (PTC): Claude writes Python code that orchestrates tool calls in a sandbox
Tool Search: Dynamic tool discovery with defer_loading for efficient context usage
MCP Integration: Tool Search combined with MCP servers via mcp_toolset
Real-World Examples: Financial data tools using yfinance
Token Savings: Up to 98% reduction in token usage for complex tasks
Technical Stack:
Anthropic Claude API (Sonnet 4.5)
Beta headers: advanced-tool-use-2025-11-20
FastMCP for MCP server implementation
Python + yfinance for financial data
ngrok for MCP server tunneling
Examples:
01_ptc_token_savings.py - Programmatic Tool Calling with token comparison
02_tool_search.py - Tool Search with 10 deferred financial tools
03_mcp_tool_search.py - MCP + Tool Search via ngrok tunnel
mcp_server.py - FastMCP server exposing financial tools
Key Concepts:
Feature
Description
Token Savings
Programmatic Tool Calling
Tool results stay in sandbox, only print() output enters context
37%
Tool Search
Only load tool definitions when discovered
85%
Combined
PTC + Tool Search together
Up to 98%
Use Cases:
Building AI agents with many tools (100+)
Reducing context window bloat from tool definitions
Processing large datasets without context overflow
MCP server integration with dynamic tool discovery
Token-efficient financial analysis agents
claude-skills/
Claude Skills API Implementation
A comprehensive implementation of Claude's Skills API for automated document generation and custom skill creation.
Key Features:
Framework: Claude Skills API with streaming support
Document Generation: PowerPoint, Excel, Word, and PDF creation
Custom Skills: Upload and manage custom skills (8MB limit)
File Management: List, download, and delete generated files
Multi-Skill Workflows: Combine multiple skills in single requests
Technical Stack:
Claude Skills API with beta features
Code execution environment (2025-08-25)
Files API (2025-04-14)
Streaming responses for real-time progress
Python SDK with uv package manager
Utilities:
list-skills.py - List all available skills
create-skill.py - Upload custom skills from directories
use-skill.py - Generate documents with single skills
multi-skill-demo.py - Complex workflows with multiple skills
A minimal Next.js starter template for building ChatKit applications with OpenAI's Agent Builder workflows.
Key Features:
Framework: Next.js with ChatKit web component
Integration: OpenAI Agent Builder workflows
Customization: Configurable themes, prompts, and UI
Session Management: Ready-to-use session endpoint
Deployment: Domain allowlist verification support
Technical Stack:
Next.js for application framework
OpenAI ChatKit web component (<openai-chatkit>)
OpenAI API integration
TypeScript for type safety
Configurable theming system
Key Components:
Session creation endpoint (/api/create-session)
ChatKit panel with event handlers
Theme and color scheme controls
Starter prompts configuration
Error overlay for debugging
Use Cases:
ChatKit application prototyping
Agent Builder workflow integration
Custom ChatKit UI development
OpenAI workflow testing
Production ChatKit deployments
mastra-overview/
Mastra Framework Exploration
An exploration of the Mastra framework for multi-LLM orchestration and agent management.
Key Features:
Framework: Mastra for multi-LLM orchestration
Multi-LLM: Support for multiple language models
Orchestration: Intelligent model selection and routing
Polyfills: Crypto polyfills for browser compatibility
Technical Stack:
Mastra framework
Multi-LLM integration
Browser compatibility polyfills
TypeScript configuration
Use Cases:
Multi-LLM agent systems
Model orchestration and routing
Framework exploration and evaluation
LLM comparison and benchmarking
smithery-example/
MCP Development Playground
A comprehensive development environment for MCP (Model Context Protocol) with FastMCP integration and testing tools.
Key Features:
Framework: Smithery + FastMCP
Development Tools: MCP playground and testing environment
Financial Integration: Example financial server implementation
Testing: Comprehensive test suite and examples
Documentation: Development guides and examples
Technical Stack:
Smithery for MCP development
FastMCP for server implementation
Testing frameworks for validation
Development tooling and playgrounds
Use Cases:
MCP server development
Protocol testing and validation
Financial data integration examples
Development environment setup
MCP learning and exploration
mcp-apps/
MCP Apps Examples (Weather + Stock Analysis)
Two minimal example MCP Apps showing how to build UI + server experiences using the MCP Apps extensions.
Key Features:
Weather App: UI + MCP server example with a simple weather workflow
Stock Analysis App: UI + MCP server example for market/stock analysis
Apps SDK: Designed to follow MCP Apps extension patterns
Docs Reference: See the MCP Apps docs for the full guide
Use Cases:
Learning MCP Apps fundamentals
Building UI-backed MCP Apps
Reference implementations for new MCP App projects
Getting Started
Each project includes comprehensive setup instructions in its respective README file. General prerequisites include:
Common Requirements
Python 3.9+ (some projects require newer; see each project README)
Valid API keys for respective services
Slack workspace access (for Slack integrations)
Environment variable configuration
Quick Start Pattern
# 1. Navigate to desired project
cd [project-name]/
# 2. Install dependencies
# Most Python projects here use uv:
uv sync
# Some projects use pip/requirements.txt:
# pip install -r requirements.txt
# 3. Configure environment
cp .env.example .env
# Edit .env with your API keys
# 4. Run the agent
# (varies by project - see individual READMEs)
Architecture Patterns
Model Context Protocol (MCP)
Multiple projects demonstrate different MCP implementation patterns:
FastMCP ASGI: Direct FastAPI integration (mcp-financial, smithery-example)
FastMCP Streamable HTTP: Modern bidirectional communication (fpl-deepagent)
Bright Data MCP: External MCP server communication
Notion MCP: Database and knowledge management integration
Agent Communication
A2A Protocol: Remote agent-to-agent communication (agent2agent)
State Management: Persistent conversation memory (bright-mcp-server-overview)