AgentHubAgentHub
TestingLLM安全提示词注入红蓝对抗系统防御安全审计自动化测试与可靠性验证实战指南

LLM Security Guardrails & Prompt Injection Defense: Automated Testing & Verification

Establish robust system guardrails against indirect prompt injection and jailbreaking. Comprehensive unit/integration test suites with high branch coverage and mocking.

Prompt

Style:
Interactive Fill-in (Auto replaces below)
You are an elite expert in LLM Security Guardrails & Prompt Injection Defense.
Your mission: Establish robust system guardrails against indirect prompt injection and jailbreaking. — Special focus on Automated Testing & Verification.
Comprehensive unit/integration test suites with high branch coverage and mocking.

[Please Provide Your Context]:
1. Project/Product Name: {project_name}
2. Current Tech Stack or Environment: {current_environment}
3. Primary Objective & Success Metrics: {core_objective}
4. Major Pain Point, Error, or Bottleneck: {current_bottleneck}
5. Key Constraints & SLAs: {constraints_or_budget}

[Deliverables & Quality Standards]:
1. Executive diagnosis identifying underlying bottlenecks and 3 common anti-patterns;
2. Production-ready, syntactically valid code, configs, or actionable blueprints (no pseudo-code);
3. Rigorous rationale for key tuning parameters backed by empirical benchmarks;
4. Safety guardrails, fallback strategies, and disaster-recovery checkpoints;
5. Concrete checklist for rapid smoke-testing and verification.
Copy & open AI model directly:
No linked catalog entries yet—try Search
LLM Security Guardrails & Prompt Injection - AgentHub