bouncer

Execute AWS CLI commands with Telegram approval for dangerous operations.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/qwer2003tw/bouncer --skill bouncer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bouncer
Source: https://github.com/qwer2003tw/bouncer/tree/main
Command: npx skills add https://github.com/qwer2003tw/bouncer --skill bouncer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill enables AI Agents to securely execute AWS commands, preventing prompt injection attacks by requiring Telegram approval for potentially dangerous operations.

Core Features & Use Cases

  • Secure AWS Command Execution: Safely run AWS CLI commands via an AI Agent.
  • Telegram Approval Workflow: Dangerous commands trigger a Telegram notification for human review and approval. Safe commands execute automatically.
  • Trust Sessions & Grants: Allows temporary automated execution for trusted scopes or specific command sets.
  • Use Case: An AI Agent needs to start an EC2 instance. Instead of executing directly, it sends a request to Bouncer. The request appears in Telegram, where a human approves it before the instance is started, preventing accidental or malicious resource changes.

Quick Start

Use the bouncer skill to execute the aws s3 ls command with the reason 'check S3 buckets'.

Frequently Asked Questions about bouncer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add human approval for AI agent AWS CLI commands?

You can add human approval for AWS CLI commands by routing AI agent execution through a Telegram approval workflow where dangerous operations trigger notifications for human review before executing.

What is Telegram approval workflow for AWS command execution?

A Telegram approval workflow for AWS command execution routes operations into blocked, safelist, or approval categories, triggering Telegram notifications for human review of potentially dangerous commands while safe commands execute automatically.

How do I classify AWS CLI commands to prevent prompt injection attacks?

You classify AWS CLI commands into blocked, safelist, and approval categories to prevent prompt injection, ensuring potentially dangerous operations require human review before an AI agent can execute them.

Can I temporarily automate AWS command execution without losing security?

Yes, you can temporarily automate AWS command execution by establishing trust sessions and grants that allow automated execution for trusted scopes or specific command sets while maintaining security boundaries.

Do I need mcporter to execute AWS commands asynchronously through an agent?

Yes, you need mcporter for local execution when running AWS commands asynchronously through an agent, as it handles the asynchronous command execution via polling.

Why does my AI agent need a bouncer for AWS operations?

An AI agent needs a bouncer for AWS operations to prevent accidental or malicious resource changes by intercepting dangerous commands and requiring human approval via Telegram before execution proceeds.