The Agent Skills Directory

[SAFE]: The skill is entirely composed of documentation and localized prompt instructions without any accompanying executable scripts or binaries.\n- [SAFE]: No tools, services, or protocols are requested in the configuration, which prevents the skill from performing unauthorized actions on the host system or network.\n- [SAFE]: Detailed instructions are provided to the AI agent to limit its scope to the current request, identify assumptions, and avoid automated web interactions.\n- [SAFE]: Multi-language support (DE, EN, ES, FR, JA, RU, ZH) is consistent and contains no hidden instructions, obfuscation, or suspicious metadata.

formbuilder-reader