89luca89

89luca89 / clampdown

Public

Run AI coding agents in hardened container sandboxes.

agent ai ai-agents claude-code containers
16
0
100% credibility
Found Mar 04, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

Clampdown runs AI coding agents inside secure isolated environments that limit file access to your project and restrict internet connections to approved services.

How It Works

1
👀 Discover Clampdown

You hear about a safe way to let AI helpers write code on your computer without risking your files or privacy.

2
📥 Get it set up

You download and prepare the tool with a simple build process that creates everything you need.

3
🔗 Link your AI service

You add your AI account details in a private spot so the helper can connect securely.

4
🚀 Start in your project

You run the tool pointing to your coding folder, and it launches the AI in a protected space.

5
🤖 AI works safely

The AI reads and edits only your project files, talks to its service online, and runs tools without accessing anything else.

✅ Code securely

Your AI helper builds your project confidently, keeping your computer safe from unexpected actions.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 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 clampdown?

Clampdown runs AI coding agents locally—like Claude Code or OpenCode—in hardened container sandboxes, locking down filesystem access to your project directory and limiting network egress to provider APIs. It solves the risk of agents executing arbitrary code via prompt injection or jailbreaks by applying the same restrictions to every tool container they spawn, such as compilers or test runners. Written in Go, it uses podman or Docker for rootless operation on Linux kernels 6.2+.

Why is it gaining traction?

It stands out by enforcing isolation outside the agent process—kernel Landlock rules, seccomp filters, and iptables chains that compromised agents can't disable—while keeping things simple with CLI commands like `clampdown claude --workdir /path/to/project`. Devs notice runtime perks like dynamic network allow/block (e.g., `clampdown network agent allow example.com`), resource limits, and opt-in credential forwarding for GitHub CLI or SSH. No more babysitting agents that could curl your SSH keys or scan your LAN.

Who should use this?

Linux devs running coding LLMs locally to generate Python scripts, HTML prototypes, or full programs without host exposure. Security teams evaluating safe ways to run coding AI locally in dev workflows. Folks debugging GitHub Actions manually or running coding models online who want offline, clamped isolation.

Verdict

Early project at 16 stars and 1.0% credibility score—maturity shows in thorough docs and tests, but expect tweaks for edge cases. Solid pick for kernel 6.2+ users needing secure local AI coding now.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.