{"skill":{"slug":"ai-security-guard","displayName":"AI Security Guard","summary":"AI安全防护系统，集成危险命令检测、多层权限模式、Hook安全机制、沙箱隔离。当用户要求安全执行命令、检测危险操作、配置权限策略、审计AI行为、保护系统安全时使用。","tags":{"hooks":"1.0.0","latest":"1.0.0","permissions":"1.0.0","sandbox":"1.0.0","security":"1.0.0"},"stats":{"comments":0,"downloads":93,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1775148693426,"updatedAt":1775148718311},"latestVersion":{"version":"1.0.0","createdAt":1775148693426,"changelog":"ai-security-guard v1.0.0\n\n- Initial release of AI安全防护系统 with core features.\n- Provides dangerous command detection, multi-layer permission modes, hook-based security mechanisms, and sandbox isolation.\n- Supports configurable permission rules at session, project, and global levels.\n- Implements critical, high, and medium risk classification for shell commands.\n- Auditing and logging for permission decisions and AI behavior.\n- Includes pre-execution security checks/checklists and sandbox execution options.","license":"MIT-0"},"metadata":null,"owner":{"handle":"e2e5g","userId":"s177nbcv26z7ksbd8gps9thykh83hgt3","displayName":"e2e5g","image":"https://avatars.githubusercontent.com/u/130967590?v=4"},"moderation":null}