skills.rest
GratefulJinx77
GratefulJinx77COMMUNITY
Explore all skills by this author →
GitHub

PromptInjection

Test AI/LLM systems for prompt injection vulnerabilities across ten attack categories.

1|Updated Mar 15, 2026
Search Tags:#security testing#AI security#penetration testing#prompt injection#LLM security
One-click install
npx skills add https://github.com/GratefulJinx77/tai --skill promptinjection-gratefuljinx77
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: PromptInjection
Source: https://github.com/GratefulJinx77/tai/tree/main/.tai/skills/security/PromptInjection
Command: npx skills add https://github.com/GratefulJinx77/tai --skill promptinjection-gratefuljinx77

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires promptfoo, garak, pyrit, browser-automation, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides a comprehensive framework for security testing of AI/LLM systems, identifying potential vulnerabilities and weaknesses.

Core Features & Use Cases

  • Attack Taxonomy: Covers 10 categories of prompt injection attacks, including direct and indirect injection, jailbreaking, system prompt extraction, and more.
  • Reconnaissance Methodology: Seven-phase process for discovering attack surfaces and identifying potential vulnerabilities.
  • Testing Tools Integration: Integrates with Promptfoo, Garak, and PyRIT for automated testing and analysis.
  • RAG-Specific Assessment: Focuses on retrieving-augmented generation systems, including document poisoning and vector database attacks.
  • Usage Examples: Offers quick and comprehensive assessment workflows for various scenarios.

Quick Start

To start testing, activate the skill and provide the target application URL. Follow the methodology steps for reconnaissance, testing, and analysis.

Frequently Asked Questions about PromptInjection

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

FAQPage Schema
How do I test my LLM application for prompt injection vulnerabilities?▼

To test for prompt injection vulnerabilities, this framework provides a seven-phase reconnaissance methodology covering direct injection, jailbreaking, and system prompt extraction using automated tools.

What types of LLM security testing methodologies are available for RAG systems?▼

RAG security testing includes assessments for document poisoning and vector database attacks, utilizing a comprehensive attack taxonomy across 10 categories of prompt injection threats.

Can I use promptfoo and Garak together for AI penetration testing?▼

Yes, you can use promptfoo, Garak, and PyRIT together for AI penetration testing, as the framework directly integrates these tools for automated vulnerability analysis.

Do I need authorization to run penetration testing on AI systems?▼

Yes, you need explicit authorization for ethical use when running penetration testing on AI systems, as the framework focuses on identifying real security weaknesses.

What is indirect prompt injection and how does it affect LLM security?▼

Indirect prompt injection is an attack vector covered in the framework's 10-category taxonomy, targeting LLM security by manipulating external content processed by the model.

Related Skills

Explore More Software Engineering Skills→
valITinovalITino
7

web-app-assessment

Automate authorized web application security assessments with active exploitation and reporting.

Community
Advanced
brucesongsbrucesongs
60

llm-red-team

Expose LLM system vulnerabilities via a six-phase red-team workflow.

Community
Advanced
CrumbgrabberCrumbgrabber
3

webapp-testing

Automate local web application testing with Python Playwright scripts.

Community
Intermediate
OWASPOWASP
141

secrets-scan

Detect hardcoded secrets in source code and configuration files.

Official
Intermediate