jnMetaCode

jnMetaCode / shellward

Public

AI Agent Security Middleware — 8-layer defense, DLP data flow, prompt injection detection, zero dependencies. SDK + OpenClaw plugin.

github.comjnMetaCodeshellward#中文 agent-security ai-agent ai-firewall ai-safety ai-security
43
4
100% credibility
Found Mar 14, 2026 at 16 stars 2x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

ShellWard is an open-source security add-on for OpenClaw AI agents that automatically hides sensitive personal data like Chinese IDs and phone numbers, blocks dangerous actions, and detects sneaky attacks.

How It Works

1
😟 Worry about your AI sharing secrets

You notice your helpful AI might accidentally show private details like phone numbers or IDs in chats.

2
🛡️ Find ShellWard protector

You discover ShellWard, a simple shield made just for AI helpers like yours, especially good with Chinese info.

3
✨ Add the shield easily

With one quick action, you turn on ShellWard to guard your AI right away.

4
💬 Use your AI normally

You keep chatting and asking your AI to do tasks, feeling more relaxed now.

5
🚫 See dangers get stopped

ShellWard quietly hides secrets, blocks risky moves, and spots tricky tricks before they happen.

6
📊 Peek at safety updates

You check simple reports to see what's protected and how safe everything is.

🎉 Safe and worry-free AI

Your AI works great while keeping all your private info hidden and secure forever.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 43 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 shellward?

Shellward is a TypeScript security plugin for OpenClaw AI agents that adds eight layers of protection against common threats like prompt injection, dangerous shell commands, and sensitive data leaks. It automatically redacts Chinese PII—ID cards with checksum validation, phone numbers across carriers, bank cards via Luhn algorithm—plus global secrets like API keys and JWTs before they reach the LLM. Install via `openclaw plugins install shellward` for zero-config enforcement, with slash commands like `/security`, `/audit`, and `/check-updates` for status and CVE scans.

Why is it gaining traction?

Unlike SecureClaw or ClawSec, which skip Chinese PII and injections, shellward handles bilingual threats with 26 rules and chain detection for exfiltration (e.g., read SSH key then POST outbound). Zero dependencies, sub-60ms redaction on 200KB text, and one-click scripts make openclaw github integration seamless—devs grab it from the openclaw github repo for instant hardening without npm bloat. Built-in 17-CVE database auto-syncs, alerting on openclaw github releases vulnerabilities.

Who should use this?

Chinese devs processing customer files in OpenClaw agents, where raw ID numbers or phones leak into chats. Security teams auditing local AI tools for compliance, or anyone chaining Bash tools who fears rm -rf slips or prompt hacks. Pairs well with openclaw github copilot workflows handling real data.

Verdict

Grab shellward from the openclaw github download if you're in China or paranoid about PII—demos and 100% test coverage shine despite 14 stars and 1.0% credibility score. Early-stage with solid docs, but watch openclaw github stars and releases for maturity; audit mode lets you test safely first.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.