The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the user to run local Python scripts for validation and generation. These scripts are included in the skill, are well-documented, and perform extensive safety checks on their inputs. They use standard libraries and common packages without invoking arbitrary shell commands or external processes.
[REMOTE_CODE_EXECUTION]: The skill requires standard Python packages (python-pptx, Pillow, lxml) to be installed via uv. These are well-known technology packages from trusted sources. There are no patterns of downloading and executing remote scripts or unverified binary payloads.
[DATA_EXFILTRATION]: The skill is explicitly designed to be network-free. Both the instructions and the code enforce local-only asset resolution and reject any external or remote relationship targets within the generated or inspected PowerPoint files.
[DYNAMIC_EXECUTION]: The code avoids eval(), exec(), and unsafe deserialization. It uses strict JSON loading (rejecting duplicate keys and non-finite numbers) and secure XML parsing (rejecting DTDs and entities to prevent XXE attacks).
[OBFUSCATION]: No obfuscation was detected. The scripts are written in clear, idiomatic Python with extensive comments and documentation. The skill includes automated checks for common placeholders and unapproved content.

pptx-posters