I will test ai agent chatbot for prompt injection jailbreak and llm security owasp llm

Your AI system may look reliable during normal use and still fail badly when a user intentionally pushes it outside its intended behavior.

I will perform an authorized AI security and red team assessment of your chatbot, RAG application, AI agent or LLM powered workflow.

The goal is to identify weaknesses such as:

PROMPT INJECTION

JAILBREAKS

SYSTEM PROMPT LEAKAGE

SENSITIVE DATA DISCLOSURE

UNSAFE TOOL USE

RAG MANIPULATION

EXCESSIVE AGENT PERMISSIONS

UNRELIABLE GUARDRAILS

UNSAFE OUTPUT BEHAVIOR

WHAT I CAN TEST

• Direct prompt injection
• Indirect prompt injection
• Jailbreak resistance
• System prompt leakage
• Instruction hierarchy failures
• Sensitive information disclosure
• Unsafe tool invocation
• Excessive agent permissions
• RAG manipulation
• Malicious document instructions
• Prompt leakage
• Context poisoning
• Agent goal hijacking
• Human approval bypass
• Unsafe function calling
• Tool permission boundaries
• Hallucination stress tests
• Output policy failures
• Authorization logic
• Memory manipulation
• Multi turn attacks
• Guardrail effectiveness
• Input validation
• Output validation
• Error handling
• Prompt security
• AI workflow security
• MCP tool safety
• Agent escalation behavior

KINDLY SEND ME A MESSAGE BEFORE ORDERING