lllons

Homemade Claude Code — AI skill for Claude Code

AI community

OUTDATED USE "TMT" INSTEAD Terminal based AI coding agent powered by Ollama.

How to install Homemade Claude Code

This entry records only its repository, not the path inside it, so there is no exact command to give. Open lllons/Homemade_claude_code and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Homemade Claude Code does

OUTDATED USE "TMT" INSTEAD Terminal based AI coding agent powered by Ollama. Create, edit and manage files with plain english commands.

Alternatives in AI

Memory Forge Rs — Stop resetting satisfying AI chats — edit the memory instead 474 ★
Codeany — Open-source AI-powered terminal agent 183 ★
Soft Ue CLI — Python CLI + UE plugin that lets Claude Code (AI coding agent) control Unreal Engine in real time 128 ★

README

OUTDATED USE "TMT" INSTEAD

"I can read, write, and execute. And if you press F2, I can listen, too." — *Local File AI Agent*

`[ Local File AI Agent ]`

A terminal-based, entirely local AI agent powered by Ollama. It doesn't just write code—it manages files, runs scripts, opens applications, and converses with you in a secure, sandboxed workspace.

`>_ WHAT IT DOES`

You type (or speak) a task in plain English. The agent figures out what files to create or modify, executes the actions, and confirms when done. It loops automatically until the task is complete.

**Example Prompts:**

"Make a file called notes.txt with my shopping list inside."
"Create a weather app with index.html, style.css, and script.js."
"Run the python script I just made."
"Open my workspace in File Explorer."

`System Capabilities`

Capability	Description
File Operations	Write, append, patch, rename, read, and delete files.
Code Execution	Automatically compiles and runs C/C++/Java, and executes Python, JS, Go, and more.
Voice Control	Press F2 to trigger local Whisper STT. The agent replies using system TTS.
App Launching	Can open permitted local apps like Notepad and File Explorer directly from the chat.
Action Batching	Performs multiple independent file operations simultaneously for faster results.

`>_ REQUIREMENTS`

Python: 3.10+
LLM Backend: Ollama running locally
Model: A compatible coding model (default: qwen2.5-coder:7b)
OS: Windows is recommended for full Voice and App Launching capabilities.

`>_ INITIALIZATION_SEQUENCE`

**0. Set Your Workspace** Open `agent.py` and modify the `ROOT_DIR` variable to dictate where the agent is allowed to see and edit files. It cannot escape this directory.

**1. Install Ollama & Pull Model** Download from [ollama.com](https://ollama.com), install it, and pull the default model: