jedi4ever

jedi4ever / addt

Public

Run AI coding agents safely in Docker containers. Your code stays isolated - no surprises on your host machine.

40
5
100% credibility
Found Feb 04, 2026 at 26 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

addt lets you run AI coding agents like Claude in isolated containers to keep your host machine safe from unintended changes.

How It Works

1
🔍 Discover safe AI coding

You hear about addt, a way to use smart AI helpers for coding without them touching your real files or computer.

2
📥 Install addt

Grab the tool with a quick download or your favorite package manager, ready in seconds.

3
⚙️ Prepare your project

Run a simple setup that learns your coding style and creates a safe playground for the AI.

4
🚀 Chat with your AI coder

Ask it to fix bugs or add features, and it works magic inside its secure bubble while seeing your code.

5
🔒 Tune protections

Adjust walls like network limits or file access to match your comfort level.

Perfect code, safe machine

Enjoy improved code from the AI, with your computer completely untouched and secure.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 26 to 40 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is addt?

addt is a Go CLI that runs AI coding agents like Claude, Codex, or Gemini safely in isolated containers using Podman, Docker, OrbStack, or Rancher Desktop. It mounts your workspace read-write at /workspace, so agents can edit files directly while blocking host access, network leaks, and surprises—no AI overwrites your system. Install via brew, then `addt run claude "Fix the bug"` to run coding agents locally with zero daemon hassle.

Why is it gaining traction?

Unlike bare agent installs, addt adds strict firewalls, scoped GitHub tokens, SSH proxies, and git hook neutralization to prevent escapes, plus persistent mode for instant restarts. Devs love the drop-in aliases (`alias claude='addt run claude'`), port forwarding for web previews, and YAML configs for teams—run coding AI locally without risking your machine. At 37 stars, it's niche but hooks security-focused users tired of unboxed LLMs.

Who should use this?

Backend devs running Python or Go agents on private repos, frontend teams building Node apps with Cursor or Copilot, or anyone cloning GitHub actions from branches needing isolation. Ideal for paranoia profiles in air-gapped setups or daily drivers wanting reproducible environments without addt-related risks like addtronic mishaps.

Verdict

Grab it if AI coding security is non-negotiable—features punch above its 1.0% credibility (low stars, early stage) with thorough docs and profiles like "paranoia." Skip for casual use; maturity lags, but it's a smart addition for local coding model runs. (187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.