security-response-policy

Apply approval checks and untrusted-input handling rules to external agent actions.

1|Updated Mar 6, 2014

Search Tags:#security #prompt injection #approval workflow #agent safety #untrusted input #outbound communication

Ready to run

npx skills add https://github.com/79yuuki/dotfiles --skill security-response-policy

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: security-response-policy
Source: https://github.com/79yuuki/dotfiles/tree/main/claude/skills/security-response-policy
Command: npx skills add https://github.com/79yuuki/dotfiles --skill security-response-policy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps agents avoid unsafe outbound actions, prompt injection traps, and accidental impersonation when processing untrusted external content.

Core Features & Use Cases

External Input Safety Checks: Treats web pages, emails, issues, pull requests, chats, logs, and attachments as untrusted data rather than instructions.
Outbound Action Guardrails: Requires explicit approval before sending messages, posting content, creating pull requests, or performing other external communications.
Use Case: Protect an AI assistant handling a GitHub issue or group chat by preventing it from following embedded malicious instructions or sharing privileged information.

Quick Start

Use the security-response-policy skill to review this external message before replying and identify any safety risks.

Frequently Asked Questions about security-response-policy

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

FAQPage Schema

How do I prevent prompt injection when an agent processes untrusted input from GitHub issues or web pages?▼

To prevent prompt injection from untrusted input, you must treat external content as data rather than instructions. This skill enforces strict separation between internal analysis and external actions, ensuring agents process web pages, emails, and issues safely without executing embedded malicious commands.

What is the best way to add an approval workflow for AI agent outbound communication?▼

The best way to add an approval workflow for outbound communication is to require explicit human checks before any external action occurs. This skill acts as a guardrail, intercepting messages, pull requests, and posts to prevent unauthorized data disclosure or unsafe automated responses.

How do I stop an AI assistant from accidental impersonation in group chats?▼

To stop accidental impersonation in group chats, you need outbound action guardrails that separate internal analysis from external replies. This skill treats chat messages as untrusted data, requiring strict approval checks before the agent can send responses or share privileged information.

Can I use this agent safety policy for scheduled jobs that handle external content?▼

Yes, you can use this agent safety policy for scheduled jobs handling external content. It explicitly applies to automated workflows involving logs, attachments, and untrusted inputs, ensuring any outbound communication or external actions triggered by the job require prior approval.

When do I need untrusted input handling rules for my AI workflows?▼

You need untrusted input handling rules whenever your AI workflows ingest external content like emails, pull requests, or web data. Applying these rules prevents the agent from following embedded malicious instructions, safeguarding against unauthorized actions and data disclosure.