clawshell

clawshell / clawshell

Public

The Runtime Security Layer for OpenClaw, the essential safety harness for PII & sensitive credentials protection.

clawshell.org ai credentials harness moltbot openclaw
127
8
100% credibility
Found Feb 17, 2026 at 32 stars 4x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

ClawShell acts as a secure middle layer for OpenClaw users, hiding real AI access codes behind virtual ones and scanning chats to block or mask personal information leaks.

How It Works

1
🔍 Discover ClawShell

You hear about ClawShell, a safety shield that keeps your AI chats in OpenClaw private and secure from accidental leaks.

2
📥 Get it set up

You add ClawShell to your computer with a simple download, ready to protect your OpenClaw experience.

3
🛡️ Run the friendly setup

With one special permission, you launch the easy guide that walks you through protecting your setup step by step.

4
Pick your AI friend
🗣️
OpenAI

Connect to the popular chatting AI for everyday conversations.

🧠
Anthropic

Link to the thoughtful AI for careful, precise responses.

5
🔒 Share secrets safely

You privately add your special access codes, and ClawShell creates safe pretend ones just for OpenClaw.

6
⚙️ Everything connects

ClawShell quietly updates OpenClaw to use the safe path and starts watching for private info.

✅ Chat securely forever

Now your OpenClaw talks are protected—no real secrets escape, and sensitive details get hidden automatically.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 32 to 127 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 clawshell?

ClawShell is a Rust-built security proxy acting as a clamshell around OpenClaw, your local LLM gateway. It swaps virtual API keys from OpenClaw for real ones stored in a locked-down `/etc/clawshell` dir, while regex-scanning requests and responses for PII like SSNs, emails, and cards—blocking risky ones or redacting before forwarding to OpenAI or Anthropic. Run `sudo clawshell onboard` for instant setup, with CLI tools for start, logs, config edits, and uninstall.

Why is it gaining traction?

It layers defense-in-depth atop OpenClaw without Vault, IdPs, or heavy deps—just Unix perms for key isolation and configurable DLP patterns. Stands out for SSE streaming passthrough, <10MB footprint via Tokio, and seamless multi-provider auth (Bearer for OpenAI, x-api-key for Anthropic). Devs grab it for quick github security scanning wins in AI pipelines, dodging credential leaks in prompts.

Who should use this?

OpenClaw teams processing user data in LLM chats, AI devs enforcing security layers of defense against PII exfil, or ops folks auditing github security advisories for credential mishandling. Suits local gateways or sidecar deploys where prompts risk leaking sensitive info.

Verdict

Early with 31 stars and 1.0% credibility score, but mature-feeling thanks to thorough docs, CLI, and tests—use it now if OpenClaw's your stack. Worth watching for broader LLM proxy adoption.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.