Extracts LLM system prompts using direct requests, jailbreak/instruction-override framing, translation/encoding tricks, and few-shot replay, combining manual payloads with automated garak and Promptfoo scanners to surface embedded secrets, routing logic, and policy leakage (OWASP LLM07:2025).
安装
npx skills add mukul975/Anthropic-Cybersecurity-Skills --skill testing-for-system-prompt-leakagenpx skills add mukul975/Anthropic-Cybersecurity-Skills --skill testing-for-system-prompt-leakage -a claude-code -gnpx skills add mukul975/Anthropic-Cybersecurity-Skills --skill testing-for-system-prompt-leakage -a codex -ggemini skills install https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git --path skills/testing-for-system-prompt-leakagegh skill install mukul975/Anthropic-Cybersecurity-Skills testing-for-system-prompt-leakagegit 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/testing-for-system-prompt-leakage ~/.claude/skills/testing-for-system-prompt-leakage # Claude Code
cp -R /tmp/mukul975__Anthropic-Cybersecurity-Skills/skills/testing-for-system-prompt-leakage ~/.agents/skills/testing-for-system-prompt-leakage # Codex, Cursor, Gemini CLI, Copilot, OpenCode