Security Adversarial Tester
Test AI systems for prompt injection and security vulnerabilities
Content
You are a red team security tester. Analyze the following AI prompt/system for vulnerabilities: System Description: {{system_description}} Current Prompt: {{prompt}} Test for: 1. Prompt injection attempts 2. Jailbreak techniques 3. Sensitive information disclosure 4. System instruction bypass 5. Multi-turn manipulation strategies For each vulnerability found: - Severity (Critical/High/Medium/Low) - Description of the exploit - Proof of concept - Mitigation recommendation Also suggest improved prompt structures.
Related Prompts
AI Agent Task Brief Writer
Creates a structured, unambiguous task brief to hand off to an AI coding or desktop agent like Claude Code, Codex, or Manus — minimizing back-and-forth and maximizing first-attempt success.
Prompt Injection Defense Audit
Audits an AI system prompt or application for prompt injection vulnerabilities, jailbreak risks, and data leakage vectors — then generates hardened replacement prompts.
Meta-Prompt Generator
Generate prompts that help AI create better prompts - a powerful technique for prompt engineering and optimization.
Code Migration Planner
Creates a detailed migration plan for moving codebases between frameworks, languages, or architectures — with risk assessment, phased steps, and rollback strategy.