Agentic Frameworks: Hands-On Guides and Tests
10 posts on agentic-framework, newest first.
An agentic framework is a toolkit for building AI agents. These are programs where a large language model runs in a loop, decides which tool to call, reads the result and keeps going until the task is done. Examples include LangGraph, CrewAI, LlamaIndex, the OpenAI Agents SDK and Google ADK. The posts below compare those frameworks, build agents from scratch in Python, and test agentic coding tools like Claude Code and DeepSeek’s dsh-superagent on real work.
Start here
| If you want to… | Read |
|---|---|
| Choose between LangGraph, CrewAI, LlamaIndex, OpenAI Agents SDK and Google ADK | Agentic Frameworks Compared: LangGraph, CrewAI & More |
| Understand the agent loop and write one yourself | What Is an AI Agent? How to Build One from Scratch in Python |
| Give an agent tools through the Model Context Protocol | What Is an MCP Server? How to Build One in Python |
| Run an agentic coding tool without constant permission prompts, safely | Claude —dangerously-skip-permissions: What It Does and When to Use It |
| See a coding agent fix a real GitHub issue end to end | I Ran DeepSeek’s dsh-superagent End to End: A Hands-On Deep Dive |
| Pick a Claude model and effort level for agent work based on measured cost | Claude Fable 5.2 Effort Levels, Tested: 26 Runs vs Opus 5.5 and Fable 5.1 |
| Self-host an agent platform and fix its storage | Understanding OpenDAL Storage in Dify: A New Year’s Journey |
Framework, coding agent or control plane?
These three terms often get mixed up. They describe different layers:
- Agentic framework. This is a library you write code against to build your own agent. It handles the loop, tool calls, state and multi-agent hand-offs. LangGraph and CrewAI are frameworks.
- Agentic AI coding tool. This is a finished agent that works inside your repository and terminal: it reads code, runs commands and edits files. Claude Code and dsh-superagent are in this group. When people search for an “agentic AI coding framework”, they usually mean one of these.
- Orchestration layer or control plane. This is the layer that decides what agents are allowed to do and keeps them governed while they run. Permission modes, allow rules, sandboxes and authorization gates live here. The posts on Claude Code permission modes and on dsh-superagent’s two-layer authorization gate cover this layer in practice.
- 24 min read
DeepSeek Harness Hands-On: What dsh Really Runs (and Sends)
A hands-on test of DeepSeek Harness (dsh 0.2.2): install it, count its 299 plugins, write your own tool, and see what each request sends to DeepSeek.
- 6 min read
Funplay MCP: Connect Claude Code to the Godot Editor (Hands-On)
Funplay MCP is an open-source MCP server for game editors. Install the Godot addon, start the server, connect Claude Code and run a first safe request.
- 7 min read
Claude Code Skills: What They Are and How to Write a SKILL.md
Claude Code skills are folders with a SKILL.md that Claude loads on demand or you run as a slash command. See where they live, the frontmatter, and a worked example.
- 8 min read
Claude Fable 5.2 Effort Levels, Tested: 26 Runs vs Opus 5.5 and Fable 5.1
Anthropic says Fable 5.2 at High matches Opus 5.5 for half the cost; an independent test found it 4.5x pricier. I ran 26 runs on three models at medium, high and xhigh, with costs from transcripts.
- 11 min read
Claude --dangerously-skip-permissions: What It Does and When to Use It
What claude --dangerously-skip-permissions really does, why it refuses to run as root, the six Claude Code permission modes, and safer ways to stop the prompts: allow rules, sandbox, auto mode.
- 27 min read
I Ran DeepSeek's dsh-superagent End to End: A Hands-On Deep Dive
A hands-on test of DeepSeek's dsh-superagent: I installed it on a Mac, had it fix a real GitHub issue end to end, and explain its two-layer authorization gate, Mandates and risks.
- 9 min read
What Is an AI Agent? How to Build One from Scratch in Python
What makes an AI agent different from a chatbot, the agent loop explained, and a working Python agent with tools built step by step on the Claude API.
- 8 min read
What Is an MCP Server? How to Build One in Python
What the Model Context Protocol is, how MCP servers expose tools, resources and prompts, and how to build, test and connect one with the Python SDK 2.x.
- 8 min read
Dify OpenDAL Storage: OPENDAL_FS_ROOT & Docker Volumes
How self-hosted Dify stores files with OpenDAL: what OPENDAL_FS_ROOT does, how it maps to Docker volumes, and how to fix Permission denied: '/Users'.
- 20 min read
TypeScript AI Agent Frameworks & Agentic Frameworks Compared
TypeScript AI agent frameworks compared: Vercel AI SDK, Mastra, LangGraph.js and OpenAI Agents SDK, with two runnable examples and a wider comparison.
FAQ
What is an agentic framework? It is a toolkit for building AI agents: programs where an LLM repeatedly plans, calls tools and acts on the results. The framework supplies the loop, tool integration and state handling, so you don’t write them yourself.
Which agentic frameworks are compared here? The comparison post covers LangGraph, LlamaIndex, the OpenAI Agents SDK, Google ADK and CrewAI. It also covers agentic AI coding tools such as Claude Code.
Do I need a framework to build an AI agent? No. “What Is an AI Agent?” builds a working agent with tools step by step in plain Python on the Claude API. That is a useful way to see what a framework does for you before you pick one.
Is Claude Code an agentic framework? Not in the library sense. Claude Code is an agentic coding tool you run in your project, not a framework you build agents with. The dangerously-skip-permissions post explains how its six permission modes control what it can do.
Is it safe to use —dangerously-skip-permissions? It removes the permission prompts entirely, so it needs care. The post explains what the flag does, why it refuses to run as root, and safer ways to stop the prompts: allow rules, a sandbox, or auto mode.