The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to generate and execute Python scripts using Numba's @njit decorator, which compiles Python code into machine code at runtime. This dynamic execution of agent-generated code represents an attack surface if the code generation logic is manipulated.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the user-provided indicator name argument.
Ingestion points: The {indicator_name} variable is derived directly from user input ($0).
Boundary markers: Absent. The indicator name is directly interpolated into file paths and Python source code without delimiters or safety warnings.
Capability inventory: The skill possesses Write, Edit, and Bash permissions, allowing it to create, modify, and execute files on the local filesystem.
Sanitization: Absent. There is no logic to sanitize the indicator_name, which could allow for path traversal (e.g., using ../../) or Python code injection if the name contains executable characters.

custom-indicator