PromptInjection - LLM Prompt Injection Security Testing
Test LLM applications for prompt injection vulnerabilities and security weaknesses
Tags
Updated: 2026-05-11Capabilities
Typical Inputs
Typical Outputs
What this skill does
- test for prompt injection
- extract system prompt
- bypass guardrails
- execute direct injection
- execute indirect injection
- perform multi-stage attacks
- conduct reconnaissance
- identify API endpoints
- inspect JavaScript
- poison RAG systems
- upload documents
- scrape web content
- manipulate API responses
- analyze DOM structure
- enumerate injection points
Inputs
- target LLM application
- written authorization
- testing scope boundaries
- application URL
- API endpoints
- document upload interface
- web scraping targets
- RAG system access
Outputs
- vulnerability findings
- security assessment report
- reconnaissance data
- attack surface map
- test results documentation
- remediation recommendations
Requirements
- written authorization for target system
- LLM application access
- browser automation capability
- security professional with offensive testing experience
- penetration testing engagement agreement