Detect and defend against indirect prompt injection hidden in web pages, documents, and images consumed by an agent, via content extraction (HTML/PDF/OCR), normalization, and scanning with LLM Guard's PromptInjection scanner or Hugging Face Prompt Guard 2. Use when an agent ingests untrusted external content and you need to screen it for injected instructions before the LLM processes it.
安装
npx skills add mukul975/Anthropic-Cybersecurity-Skills --skill detecting-indirect-prompt-injectionnpx skills add mukul975/Anthropic-Cybersecurity-Skills --skill detecting-indirect-prompt-injection -a claude-code -gnpx skills add mukul975/Anthropic-Cybersecurity-Skills --skill detecting-indirect-prompt-injection -a codex -ggemini skills install https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git --path skills/detecting-indirect-prompt-injectiongh skill install mukul975/Anthropic-Cybersecurity-Skills detecting-indirect-prompt-injectiongit clone --depth 1 https://github.com/mukul975/Anthropic-Cybersecurity-Skills /tmp/mukul975__Anthropic-Cybersecurity-Skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/mukul975__Anthropic-Cybersecurity-Skills/skills/detecting-indirect-prompt-injection ~/.claude/skills/detecting-indirect-prompt-injection # Claude Code
cp -R /tmp/mukul975__Anthropic-Cybersecurity-Skills/skills/detecting-indirect-prompt-injection ~/.agents/skills/detecting-indirect-prompt-injection # Codex, Cursor, Gemini CLI, Copilot, OpenCode