Security advisory — Malicious litellm versions 1.82.7 and 1.82.8 were removed from PyPI (potential API key exfiltration). Uninstall them, rotate exposed credentials, and upgrade to a safe release (e.g. 1.82.9+ per upstream). Run pip show litellm to verify. PyPI · README

Building Your First Agent

Step-by-step guide to building a research agent.

Building Your First Agent

This guide walks through building a real-world research agent step by step.

What We're Building

A research agent that:

  1. Accepts a research topic
  2. Searches the web for information
  3. Synthesizes findings into a structured report

Step 1: Set Up

bash
1pip install agenticx
2export OPENAI_API_KEY="your-key"

Step 2: Define Your Tools

python
1# tools.py
2from agenticx.tools import tool
3import httpx
4

5@tool
6def search_web(query: str) -> str:
7    """Search the web for information about a topic.
8    
9    Args:
10        query: The search query
11    
12    Returns:
13        Search results as text
14    """
15    # Replace with your preferred search API
16    response = httpx.get(
17        "https://api.search.com/search",
18        params={"q": query, "key": "your-api-key"}
19    )
20    return response.text
21

22@tool
23def fetch_page(url: str) -> str:
24    """Fetch the content of a web page.
25    
26    Args:
27        url: The URL to fetch
28    
29    Returns:
30        Page content as text
31    """
32    response = httpx.get(url, follow_redirects=True)
33    return response.text[:5000]  # First 5000 chars

Step 3: Define the Agent

python
1# agent.py
2from agenticx import Agent
3

4research_agent = Agent(
5    id="research-agent",
6    name="Research Assistant",
7    role="Senior Research Analyst",
8    goal=(
9        "Conduct thorough research on any given topic. "
10        "Find authoritative sources, synthesize information, "
11        "and produce clear, well-structured reports."
12    ),
13    backstory=(
14        "You are an expert researcher with a background in "
15        "information synthesis and critical analysis."
16    ),
17    organization_id="my-research-org",
18    max_iter=15,
19    verbose=True
20)

Step 4: Create and Run a Task

python
1# main.py
2from agenticx import Task, AgentExecutor
3from agenticx.llms import OpenAIProvider
4from tools import search_web, fetch_page
5from agent import research_agent
6

7def research(topic: str) -> str:
8    task = Task(
9        id="research-task",
10        description=f"Research: {topic}",
11        expected_output=(
12            "A structured research report with:\n"
13            "1. Executive summary\n"
14            "2. Key findings\n"
15            "3. Detailed analysis\n"
16            "4. Sources and references"
17        )
18    )
19    
20    llm = OpenAIProvider(model="gpt-4o")
21    executor = AgentExecutor(
22        agent=research_agent,
23        llm=llm,
24        tools=[search_web, fetch_page]
25    )
26    
27    return executor.run(task)
28

29if __name__ == "__main__":
30    result = research("Multi-agent AI systems impact on software development")
31    print(result)

Step 5: Run It

bash
1python main.py

Enhancements

Add Memory

python
1from agenticx.memory import MemoryManager
2

3memory = MemoryManager()
4executor = AgentExecutor(agent=research_agent, llm=llm, tools=[...], memory=memory)

Add Observability

python
1from agenticx.observability import ConsoleTracer
2

3tracer = ConsoleTracer()
4executor = AgentExecutor(agent=research_agent, llm=llm, tools=[...], tracer=tracer)

Use the CLI

bash
1agx project create research-bot --template basic
2cd research-bot
3agx run agent.py --verbose

Next Steps