greshake/llm-security

• highreadme#1

  Reposition README opening to clarify its role as a research/demonstration resource

  原因：

  当前

  ## New: Demonstrating Indirect Injection attacks on Bing Chat

  复制粘贴的修复

  复制

  ## LLM Security: Demonstrations and Research on Indirect Prompt Injection
  This repository provides a comprehensive collection of proof-of-concept demonstrations and research findings on LLM security vulnerabilities, with a focus on indirect prompt injection. It serves as an educational resource for security researchers, developers, and practitioners to understand and explore new attack vectors against app-integrated Large Language Models, accompanying our detailed paper on ArXiv.

  待办完成忽略
• hightopics#2

  Add specific topics to improve categorization

  原因：

  复制粘贴的修复

  复制

  llm-security, prompt-injection, llm-vulnerabilities, ai-security, large-language-models, security-research, proof-of-concept

  待办完成忽略
• mediumhomepage#3

  Add a homepage URL linking to the research paper

  原因：

  复制粘贴的修复

  复制

  https://arxiv.org/pdf/2303.06572.pdf

  待办完成忽略

本次扫描解析到的品类 GEO 通道：google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• 品类问题

  How to identify and mitigate prompt injection vulnerabilities in my AI applications?

  你：未被推荐
  AI 推荐顺序：

  1. Garak (llm-random-walk/garak)
  2. LLM Guard (laiyer-ai/llm-guard)
  3. Lakera Guard
  4. Rebuff (protectai/rebuff)
  5. Prompt Security
  6. NeMo Guardrails (NVIDIA/NeMo-Guardrails)
  7. LangChain (langchain-ai/langchain)

  AI 推荐了 7 个替代方案，却始终没点名 greshake/llm-security。这就是要补上的差距。

  查看 AI 完整回答
• 品类问题

  What are best practices for securing language models integrated into applications?

  你：未被推荐
  AI 推荐顺序：

  1. OWASP ESAPI
  2. Sanitizer
  3. DOMPurify (cure53/DOMPurify)
  4. OpenAI Moderation API
  5. Google Cloud Content Moderation API
  6. Azure Content Moderator
  7. Auth0
  8. Okta
  9. AWS IAM
  10. Cloudflare
  11. NGINX
  12. Kong Gateway (Kong/kong)
  13. Presidio (microsoft/presidio)
  14. Google Cloud Data Loss Prevention (DLP) API
  15. AWS Macie
  16. Datadog
  17. Prometheus (prometheus/prometheus)
  18. Grafana (grafana/grafana)
  19. Kubernetes (kubernetes/kubernetes)
  20. HashiCorp Vault (hashicorp/vault)
  21. Tenable Nessus

  AI 推荐了 21 个替代方案，却始终没点名 greshake/llm-security。这就是要补上的差距。

  查看 AI 完整回答

• Metadata completeness

  warn

  建议：

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of greshake/llm-security?

  pass
  ✓AI 明确点名了 greshake/llm-security

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• If a team adopts greshake/llm-security in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 greshake/llm-security

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• In one sentence, what problem does the repo greshake/llm-security solve, and who is the primary audience?

  pass
  ✓AI 明确点名了 greshake/llm-security

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？