The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates and saves a Python script named {symbol}_strategy_comparison.py to the local file system. The agent is then instructed to execute this dynamically generated code to perform the strategy comparison and generate visualizations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided symbols and strategy names via $ARGUMENTS and interpolates them directly into file paths and the content of the generated script, which represents an attack surface for indirect injection.
Ingestion points: User-provided input through the $ARGUMENTS variable, specifically the $0 (symbol) and remaining strategy name arguments.
Boundary markers: The instructions do not define any boundary markers or delimiters to isolate user-provided data from the generated script logic.
Capability inventory: The skill utilizes the Write tool to create directories and files, and the Bash tool for script execution and file system operations.
Sanitization: No explicit sanitization, validation, or escaping of the user-supplied arguments is performed before they are used to construct filenames and code content.

strategy-compare