The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): Utility scripts in the scripts/ directory facilitate skill management through standard filesystem operations like directory traversal and zip archival. The quick_validate.py script specifically uses yaml.safe_load() to ensure that metadata parsing does not lead to unsafe deserialization or execution of arbitrary code.- [PROMPT_INJECTION] (SAFE): Reference documentation in the references/ folder contains structural templates and workflow examples. These serve as benign formatting guidelines for the agent and do not include any markers, bypass attempts, or instructions designed to override system safety protocols.

skill-creator