Skip to content

README.md

Latest commit

 

History

History
78 lines (64 loc) · 5.02 KB

README.md

File metadata and controls

78 lines (64 loc) · 5.02 KB
Raw

Python Examples

This directory contains runnable Python examples showing how to integrate Agent Assembly with popular AI agent frameworks.

What lives here

Sub-project Framework What it demonstrates
langchain-basic-agent/ LangChain Wire Agent Assembly SDK into a basic LangChain agent
openai-agents-sdk/ OpenAI Agents SDK Govern tool calls made by an OpenAI Agents SDK agent
llamaindex-tool-policy/ LlamaIndex Enforce tool-level allow/deny policies with LlamaIndex
custom-tool-policy/ — Write a custom Python tool wrapper with SDK
langchain-research-agent/ LangChain ReAct research agent with budget, egress allowlist, and credential-leak blocking
crewai-research-crew/ CrewAI Multi-agent crew with delegation tracking, file-write approval, and shared budget
langgraph/ LangGraph Node-level governance on a compiled StateGraph, blocking a destructive tool mid-graph
pydantic-ai/ Pydantic AI Tool-call governance driven offline by TestModel (allow / deny / pending)
google-adk/ Google ADK Scripted offline tool trajectory governing BaseTool.run_async (no cloud creds)
haystack-tool-policy/ Haystack Govern a real Haystack agent via the native adapter — real Tool.invoke allow/deny through a ToolInvoker
smolagents-tool-policy/ Smolagents Govern real smolagents.Tool calls via Tool.__call__, blocking a destructive tool offline (no model creds)
microsoft-agent-framework-tool-policy/ Microsoft Agent Framework Govern FunctionTool.invoke (allow / deny / pending); mock + live paths
agno-tool-policy/ Agno Govern real Agno FunctionCall.execute tool calls (allow / deny / pending); offline, no model creds
autogen-tool-policy/ AutoGen / ag2 Govern real autogen_core.tools.FunctionTool.run_json calls, blocking a destructive tool offline (no model creds)
semantic-kernel-tool-policy/ Semantic Kernel Govern real Semantic Kernel Kernel.invoke calls, blocking a destructive tool offline (no model creds)
strands-agents-tool-policy/ Strands Agents Govern real @strands.tool invocations, blocking a destructive tool offline (no model creds)

All examples use the agent-assembly Python package (available on PyPI).

Framework coverage goal

Decision (AAASM-4458): this repo adopts an explicit framework-coverage matrix — the table below — instead of relying only on the descriptive catalog above. The catalog grows organically and makes gaps invisible; a stated matrix makes any missing framework visible by inspection, so the next gap surfaces without another ad-hoc audit. When a new framework gains an example, move it to Have example; when the team commits to one, add it under Planned; frameworks intentionally out of scope go under Not planned with a one-line reason.

Framework Status Notes
LangChain Have example langchain-basic-agent, langchain-research-agent
LangGraph Have example langgraph
CrewAI Have example crewai-research-crew
OpenAI Agents SDK Have example openai-agents-sdk
Pydantic AI Have example pydantic-ai
Google ADK Have example google-adk
LlamaIndex Have example llamaindex-tool-policy
Haystack Have example haystack-tool-policy
Smolagents (Hugging Face) Have example smolagents-tool-policy
Microsoft Agent Framework Have example microsoft-agent-framework-tool-policy (distinct from Semantic Kernel)
Agno Have example agno-tool-policy
AutoGen / ag2 Have example autogen-tool-policy
Microsoft Semantic Kernel Have example semantic-kernel-tool-policy (distinct from the Agent Framework)
Strands Agents (AWS) Have example strands-agents-tool-policy
Framework-agnostic (plain Python) Have example custom-tool-policy
(next candidate) Not planned yet Open a ticket to move a framework to Planned before adding an example

Prerequisites

  • Python >= 3.12
  • uv (recommended) or pip
  • A running Agent Assembly gateway (see the sub-project README for local dev options)

Expected sub-project structure

Each sub-project in this directory should follow this layout:

python/<example-name>/
  README.md         ← prerequisites, setup, run, expected output, troubleshooting
  .env.example      ← template for any required secrets or config (never .env)
  pyproject.toml    ← or requirements.txt
  src/
    main.py         ← entrypoint

Back to root

← Agent Assembly Examples