I will perform ai model jailbreak, prompt injection testing, and llm security analysis

I will perform a complete AI model security evaluation, including jailbreak testing, prompt injection analysis, safety filter bypass attempts, and vulnerability detection for LLMs and chatbots.

I specialize in identifying weaknesses in AI systems and exposing how attackers can manipulate your model to produce unsafe, harmful, biased, or unauthorized responses.

WHAT I WILL TEST

1. Jailbreak Attacks

• DAN-style jailbreaks
• System prompt override
• Personality switching attacks
• Role escape and forced internal mode activation

2. Prompt Injection Attacks

• Direct prompt manipulation
• Indirect prompt injection (hidden instructions)
• HTML/Markdown prompt injection
• Context poisoning attacks

3. Safety Filter Bypass

• Bypassing safety guardrails
• Getting model to output restricted content
• Ethical boundaries bypass
• Hidden policy conflicts

4. Multi-Step Adversarial Attacks

• Chain-of-thought manipulation
• Multi-layer contextual attacks
• Recursive instruction loopholes

5. Hallucination & Bias Testing

• Trigger conditions for hallucination
• Bias amplification checks
• Contradictory responses
• Incorrect reasoning traps