Skills
.Work. You
Rest

llm-output-safety-reviewer

Community

Stop unsafe AI output before it becomes a breach.

Software Engineering#xss#injection#sandboxing#ssrf#llm security#output handling#second-order
Authornguyenpv1980-wq
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill reviews how applications handle LLM-generated output so unsafe content does not become XSS, injection, SSRF, RCE, or second-order security issues.

Core Features & Use Cases

  • Output Sink Review: Traces model output into HTML, markdown, SQL, shell commands, file paths, URLs, tool arguments, and stored content.
  • Safe Rendering and Execution Checks: Verifies context-correct escaping, allowlisting, parameterization, and sandboxing for generated code execution.
  • Second-Order Risk Analysis: Detects when stored model output is later re-used as trusted data in admin consoles, exports, or downstream services.
  • Use Case: A team adds an AI dashboard that renders assistant replies as markdown and runs generated SQL, and this Skill identifies the exact unsafe sinks and required fixes.

Quick Start

Review our AI feature's output handling and list every place model output is rendered, executed, stored, or used to drive an action.

Dependency Matrix

Required Modules

None required

Components

references

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: llm-output-safety-reviewer
Download link: https://github.com/nguyenpv1980-wq/Project-Aegis/archive/main.zip#llm-output-safety-reviewer

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 545,000+ vetted skills library on demand.