The Agent Skills Directory

[Dynamic Code Execution]: The skill incorporates tools such as execute_python_script and visualize that enable the agent to generate and run Python code for data transformation and statistical analysis. This is a core functionality and is accompanied by instructions that limit the environment to specific libraries while explicitly prohibiting network and unauthorized file system access.
[Data Ingestion and Analysis]: The agent is designed to process tables within a workspace. This capability involves reading schema information and sample data to inform its actions. While processing external data is a standard vector for indirect prompt injection, the skill includes guidelines for neutral interpretation and structured reporting to maintain analytical integrity.
[Human-in-the-Loop Interaction]: Through the ask_user action, the skill provides a mechanism for the agent to pause and request clarification or choices from the user. This design supports controlled execution and ensures the agent remains aligned with user intent during complex tasks.

core