AI Skills & Agents

AgentShield Security Scanner

Protect your repository from leaked API keys, prompt injection attacks, and dangerous hook execution.

4-Layer Security Engine

AgentShield is designed to protect autonomous AI coding loops from compromising your operating system, private keys, or codebase integrity.

1. Secret Leak Scanner

Scans for raw OpenAI, Anthropic, AWS, GitHub PAT, and DB URIs.

2. Prompt Injection Defense

Audits markdown instructions from jailbreak and override attacks.

3. Malicious Hook Audit

Flags destructive script execution like rm -rf without confirmation.

4. MCP Permission Health

Audits over-privileged MCP server arguments.

bash

npx awesome-ai-tools scan

Was this page helpful?