Resources
Guides for people who run AI agents
How skills, automations and harnesses change what your agent hands back, plus profiles of the open-source tools behind them.
What an AI Agent Skill Is, and Why the Same Request Comes Back Better
A skill is the written method you would hand a sharp new colleague. What one holds, how an agent picks it up, and why the same request comes back better with one attached.
How to Use Skills in Claude Code: Add One, Name It, Check It Ran
A plain guide to skills in Claude Code: where a skill goes, how to write the line that lets Claude find it, how to call it by name, and how to tell it was followed.
Prompts vs Skills: Why Rewriting the Request a Third Time Doesn't Work
A prompt asks for the work. A skill carries the steps, checks and traps behind it. Why rewording the same request rarely fixes a generic draft, and what to move out of the prompt instead.
Six Skills to Try First if You Use an Agent for Client Work
Six everyday client jobs, from the proposal to the Friday update: where an agent usually goes wrong on each, what a written method changes, and what to check before it goes out.
Make Your Agent Check Its Sources Before You Do
AI agents can hand back references that look right and don't exist. How to give yours a research method: three sources per claim, every reference checked, and a record of where each claim came from.
Turn a Weekly Job Into an Automation Your Agent Runs While You're Away
A weekly job your agent can finish on its own: one skill per step, a check between steps that stops the run when something is off, and a finished draft for you to review.
What an Agent Harness Is, and When to Start From a Ready-Made One
A harness is everything wired up around the model so it can do real work. What one holds, why it changes results, and when starting from one that already runs beats building your own.
One Folder, One Standard: Giving Your Whole Team's Agents the Same Methods
When everyone on a team keeps private prompts, the work comes back five different ways. How to keep one shared set of methods, who should own them, and how changes reach every agent.
Aider: pair programming in the terminal
An open-source terminal tool that edits your code, runs tests, and commits to git, working with Claude, GPT and local models.
AutoGen: Microsoft's multi-agent conversation framework
AutoGen lets you build teams of AI agents that talk to each other and to tools. Here is what it does and where it now sits in Microsoft's agent stack.
AutoGPT: the repo that started the autonomous agent wave
AutoGPT popularized the plan, act, observe loop in 2023. In 2026 it is a full agent platform with a paid hosted tier and a free self-hosted path.
Browser Use: agents that drive a real browser
Browser Use is an open-source Python library that lets an AI agent click, type, and navigate a real browser from natural-language instructions.
Cline: an autonomous coding agent inside VS Code
An open-source agent that edits code and runs commands inside your editor, asking for approval on every step by default.
CrewAI: role-based agent crews
CrewAI organizes agents by role instead of by graph, pairing autonomous Crews with event-driven Flows for tighter control.
Dify: a visual builder for agents and workflows
An open-source platform for building AI agents and workflows on a drag-and-drop canvas, with a built-in RAG pipeline and model management.
Goose: Block's extensible local agent
A general-purpose agent that runs on your own machine as a desktop app, CLI, or API, connecting to 70+ tools over MCP and 15+ model providers.
LangGraph: building agents as graphs
LangGraph models an agent as nodes and edges instead of a single prompt loop, giving you durable execution, memory, and human review points.
MetaGPT: an agent team that works like a software company
MetaGPT assigns product manager, architect and engineer roles to a group of agents so one requirement can turn into a working project, not just a code snippet.
n8n: workflow automation with AI agent nodes
n8n is a fair-code workflow automation platform that wires AI agents, tools, and human approvals into visual workflows you can self-host.
Open Interpreter: running code on your own machine
Open Interpreter is now a Rust-based coding agent forked from OpenAI's Codex, tuned for low-cost models, that edits and runs code in your terminal.
OpenAI Agents SDK: handoffs, guardrails and tracing
OpenAI's production framework for multi-agent workflows, built around agents, handoffs, guardrails and sessions, with tracing to see what actually ran.
OpenHands: an open-source coding agent
OpenHands is a self-hosted control center for coding agents, with a sandboxed runtime and support for third-party agents like Claude Code and Codex.
smolagents: Hugging Face's minimal agent library
smolagents is Hugging Face's small Python library for building AI agents that write their actions as executable code instead of JSON tool calls.