skills.rest
EntroVyx
EntroVyxCOMMUNITY
Explore all skills by this author →
GitHub

llm-prompt-injection

Test LLM endpoints for prompt injection vulnerabilities using Python scripts.

3|1|Updated Jul 2, 2026
Search Tags:#security#ai#llm#jailbreak#prompt-injection#redteam
One-click install
npx skills add https://github.com/EntroVyx/hermes-agent-offsec --skill llm-prompt-injection-entrovyx
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-prompt-injection
Source: https://github.com/EntroVyx/hermes-agent-offsec/tree/main/skills/offsec/recon/llm-prompt-injection
Command: npx skills add https://github.com/EntroVyx/hermes-agent-offsec --skill llm-prompt-injection-entrovyx

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests.

What problem does it solve?

This skill addresses the critical security challenge of prompt injection in LLM applications, providing a comprehensive catalog of techniques to identify and defend against vulnerabilities where untrusted data overrides system instructions.

Core Features & Use Cases

  • Technique Catalog: Access 40+ tested attack vectors including direct, indirect, and agentic injection methods.
  • Defense Mapping: Implement multi-layer security architectures like the Silversky model and Guardian model patterns.
  • Use Case: Security researchers can use this skill to audit hardened GPT-4o-class deployments for vulnerabilities in RAG pipelines, agentic tool-call chains, and structured data interpolation.

Quick Start

Use the llm-prompt-injection skill to run a baseline context length analysis and boolean extraction test against your target endpoint.

Frequently Asked Questions about llm-prompt-injection

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

FAQPage Schema
How do I test prompt injection vulnerabilities in LLM applications?▼

To test prompt injection vulnerabilities, you can run security assessment scripts against your target AI endpoints. This skill evaluates security posture across RAG pipelines, multi-modal inputs, and structured data fields using a catalog of direct, indirect, and agentic attack techniques.

What is indirect prompt injection and how does it affect RAG pipelines?▼

Indirect prompt injection occurs when untrusted data overrides system instructions through retrieved content. This skill identifies these vulnerabilities by testing RAG pipelines and structured data interpolation with 40+ cataloged attack vectors.

Can I audit GPT-4o class deployments for system prompt extraction?▼

Yes, you can audit hardened GPT-4o-class deployments for system prompt contents. The skill executes boolean extraction tests and context length analysis to map system prompts and identify security weaknesses.

Do I need Python to run LLM red team security assessments?▼

Yes, you need Python 3 and the requests library to execute the deterministic security assessment scripts. These dependencies are required to run the attack techniques against your target AI endpoints.

What's the best way to defend against agentic prompt injection in tool-call chains?▼

To defend against agentic prompt injection, implement multi-layer security architectures like the Silbersky model and Guardian model patterns. This skill provides defense mapping to secure agentic tool-call chains against detected vulnerabilities.

Related Skills

Explore More Software Engineering Skills→
SoljournerSoljourner
45

get-available-resources

Enumerate installed Python packages, databases, and software tools with versions and connectivity.

Community
Basic
cdman28cdman28

phase-7-seo-security

Implement SEO meta tags, structured data, and security measures against XSS and CSRF.

Community
Advanced
PrepLabsAIPrepLabsAI
94

prompt-engineering-interviewer

Assess AI prompt engineering proficiency through multi-phase technical interviews.

Official
Advanced
ihatesea69ihatesea69
3

nemo-guardrails

Implement runtime safety for LLM applications with NVIDIA NeMo Guardrails.

Community
Advanced