原始内容
📦 Install Now · 🤖 Explore Agents · 📖 Documentation · ⭐ Star this repo
What is this?
AI Agents Assistant is a Claude Code plugin with 7 production-grade agents for building AI agent systems using LangChain, LangGraph, Anthropic, and OpenAI patterns.
📑 Table of Contents
Click to expand
🚀 Quick Start
Prerequisites
- Claude Code CLI v2.0.27+
- Active Claude subscription
Installation (Choose One)
Option 1: From Marketplace (Recommended)
# Step 1️⃣ Add the marketplace
/plugin marketplace add pluginagentmarketplace/custom-plugin-ai-agents
# Step 2️⃣ Install the plugin
/plugin install custom-plugin-ai-agents@pluginagentmarketplace-ai-agents
# Step 3️⃣ Restart Claude Code
# Close and reopen your terminal/IDE
Option 2: Local Installation
# Clone the repository
git clone https://github.com/pluginagentmarketplace/custom-plugin-ai-agents.git
cd custom-plugin-ai-agents
# Load locally
/plugin load .
# Restart Claude Code
✅ Verify Installation
After restart, you should see these agents:
custom-plugin-ai-agents:01-ai-agent-fundamentals
custom-plugin-ai-agents:02-llm-integration
custom-plugin-ai-agents:03-rag-systems
custom-plugin-ai-agents:04-tool-calling
custom-plugin-ai-agents:05-multi-agent
custom-plugin-ai-agents:06-agent-memory
custom-plugin-ai-agents:07-agent-safety
✨ Features
| Feature | Description |
|---|---|
| 🤖 7 Agents | Production-grade AI agent specialists |
| 🛠️ 7 Skills | Atomic capabilities with code examples |
| ⌨️ 5 Commands | Quick slash commands for learning |
| 🔄 SASMP v1.3.0 | Full protocol compliance |
| 📚 2024-2025 Patterns | LangGraph, Claude Tool Use, OpenAI Functions |
🤖 Agents
7 Production-Grade Agents
| # | Agent | Expertise |
|---|---|---|
| 1 | AI Agent Fundamentals | ReAct patterns, cognitive loops, architectures |
| 2 | LLM Integration | API orchestration, prompting, cost optimization |
| 3 | RAG Systems | Embeddings, chunking, hybrid search, reranking |
| 4 | Tool Calling | Function calling, schemas, validation |
| 5 | Multi-Agent | Orchestrator-worker, hierarchical systems |
| 6 | Agent Memory | Short/long-term, semantic retrieval |
| 7 | Agent Safety | Guardrails, filtering, compliance |
Agent Capabilities
Each agent includes:
- ✅ Clear role & responsibility boundaries
- ✅ Input/Output schemas (TypeScript)
- ✅ Error handling patterns
- ✅ Fallback strategies
- ✅ Token/cost optimization configs
- ✅ Troubleshooting decision trees
- ✅ Production code examples
🛠️ Skills
7 Atomic Skills
| Skill | Bonded Agent | Purpose |
|---|---|---|
ai-agent-basics |
01-ai-agent-fundamentals | Build ReAct agents with LangGraph |
llm-integration |
02-llm-integration | Integrate Claude, OpenAI, local models |
rag-systems |
03-rag-systems | Build production RAG pipelines |
tool-calling |
04-tool-calling | Implement function calling |
multi-agent |
05-multi-agent | Build orchestrator-worker systems |
agent-memory |
06-agent-memory | Add memory to agents |
agent-safety |
07-agent-safety | Implement safety guardrails |
Invoke Skills
# Example: Invoke the RAG skill
Skill("custom-plugin-ai-agents:rag-systems")
⌨️ Commands
| Command | Description | Usage |
|---|---|---|
/learn |
Start your AI agent journey | /learn |
/roadmap |
View learning roadmap | /roadmap ai-agents |
/assess |
Test your knowledge | /assess tool-calling intermediate |
/resources |
Get learning resources | /resources rag-systems tutorials |
/project |
Find project ideas | /project multi-agent advanced |
📋 Best Practices (2024-2025)
From LangChain/LangGraph
- Use LangGraph for new agent implementations
- Implement human-in-the-loop checkpoints
- Add observability with LangSmith
From Anthropic
- One job per subagent
- Orchestrator handles global planning
- Keep tool permissions narrow
From OpenAI
- Enable strict mode for function calling
- Use structured outputs
- Implement circuit breakers
📁 Project Structure
custom-plugin-ai-agents/
├── 📁 .claude-plugin/
│ ├── plugin.json
│ └── marketplace.json
├── 📁 agents/ # 7 production agents
│ ├── 01-ai-agent-fundamentals.md
│ ├── 02-llm-integration.md
│ ├── 03-rag-systems.md
│ ├── 04-tool-calling.md
│ ├── 05-multi-agent.md
│ ├── 06-agent-memory.md
│ └── 07-agent-safety.md
├── 📁 skills/ # 7 atomic skills
├── 📁 commands/ # 5 slash commands
├── 📁 hooks/
├── 📄 README.md
├── 📄 CHANGELOG.md
└── 📄 LICENSE
📚 Documentation
| Document | Description |
|---|---|
| CHANGELOG.md | Version history |
| CONTRIBUTING.md | How to contribute |
| LICENSE | License information |
External Resources
| Resource | Link |
|---|---|
| LangGraph Docs | https://langchain-ai.github.io/langgraph/ |
| Anthropic Docs | https://docs.anthropic.com/ |
| OpenAI Docs | https://platform.openai.com/docs |
| OWASP LLM Top 10 | https://owasp.org/www-project-top-10-for-large-language-model-applications/ |
📅 Metadata
| Field | Value |
|---|---|
| Version | 2.0.0 |
| Last Updated | 2025-01-01 |
| Status | Production Ready |
| SASMP | v1.3.0 |
| Agents | 7 |
| Skills | 7 |
| Commands | 5 |
🤝 Contributing
Contributions are welcome! Please read our Contributing Guide.
- Fork the repository
- Create your feature branch
- Follow the production patterns in existing agents
- Submit a pull request
⚠️ Security
Important: This repository contains patterns for AI agent development.
- ✅ Always implement safety guardrails
- ✅ Follow OWASP LLM Top 10 guidelines
- ✅ Test for prompt injection vulnerabilities
- ✅ Report security issues privately via Issues
📝 License
Copyright © 2025 Dr. Umit Kacar & Muhsin Elcicek
Custom License - See LICENSE for details.
👥 Contributors
|
Dr. Umit Kacar Senior AI Researcher & Engineer |
Muhsin Elcicek Senior Software Architect |
Made with ❤️ for the Claude Code Community
