What are AI agents?
Find the full "AI Agents for Beginners" Course and code samples here ➡️ aka.ms/ai-agents-beginners
In this lesson we ask, what are AI agents? We then discuss their best use cases and what you need to start building them.
In this video you'll find: 00:00 Introduction 00:35 Parts of an AI agent 02:22 Code sample and demonstration
GitHub AI Trend Weekly: June 29 - July 5, 2026: agency-agents, Strix, OmniRoute Featured
This week we look at the Top 10 new Stars for GitHub AI / Agent / LLM related projects from June 29 to July 5, 2026. agency-agents, Strix, codebase-memory-mcp, OpenMontage continue to dominate the front ranks. Among the new entries, focus on ai-berkshire, OmniRoute, video-use, and herdr: one for investment research workflow, one as a model gateway, one that hands video editing to a coding agent, and one that orchestrates multiple agents in the terminal. Project URLs: agency-agents: https://github.com/msitarzewski/agency-agents Strix: https://github.com/usestrix/strix codebase-memory-mcp: https://github.com/DeusData/codebase-memory-mcp OpenMontage: https://github.com/calesthio/OpenMontage ai-berkshire: https://github.com/xbtlin/ai-berkshire OmniRoute: https://github.com/d...
This open-source project lets AI generate editable CAD source code with one sentence, and can export STEP and URDF
Mechanical design is stuck again.
Using AI to generate a CAD model, but what you get is an STL mesh, and you can't change the hole positions.
To export the data as STEP format for factory use, you need to rebuild the geometry.
Robot description files require manually adding a lot of XML tags, which is error-prone.
Recently, I came across a project on GitHub called text-to-cad, which can generate CAD source code and engineering files with one click. These problems finally have hope.
It is an Agent-oriented CAD skill library with over 7400 stars on GitHub.
In a nutshell:
Use AI to write editable CAD source code and export it as engineering files like STEP, URDF.
Let me show you how convenient this thing is.
I want to design a robot chassis.
I casually opened ClaudeCode and said, "Generate a rectangular base with four mounting holes and two motor brackets."
It started writing...
AI agents are easy to build now. Managing and trusting them is the real problem.
Everyone is building AI agents now.
n8n agents. Cursor agents. Claude Code agents. MCP setups. LangChain scripts. Personal assistants. Client automations.
But nobody has a clean place to answer the basic questions:
What agents do I have? What do they do? What tools and data can they access? Which ones are dangerous? Which ones are tested? Which ones are wasting money? Which ones are ready to share, sell or deploy?
That’s the gap.
Right now agents live across n8n workflows, GitHub repos, Cursor projects, Claude chats, Notion docs and random screenshots.
It’s messy. And as agents get more powerful, messy becomes risky.
So I’m working on an Agent Control Center.
One dashboard where every AI agent gets an Agent Passport:
purpose tools data access permissions risks test checklist cost estimate version/status shareable docs
For hobby builders, it’s a clean overview of all your agents...