The Agent Skills Directory

[COMMAND_EXECUTION]: The tools generate_chart and create_visual_design defined in SKILL.md are designed to execute arbitrary Python code provided via the python_code string parameter. This allows for powerful visualization logic but creates a vector for command execution if the agent is misled during code generation.
[REMOTE_CODE_EXECUTION]: The skill facilitates the runtime generation and execution of scripts. It lists several libraries for this purpose, including matplotlib, plotly, bokeh, reportlab, fpdf, Pillow, Wand, and opencv-python.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it transforms untrusted user requirements into executable code without explicit sanitization or boundary markers.
Ingestion points: User-provided data or requests for visual designs (SKILL.md).
Boundary markers: Absent; there are no instructions to isolate or delimit user-provided data when generating the python_code string.
Capability inventory: The skill possesses both file-writing and arbitrary code execution capabilities through the defined tools (SKILL.md).
Sanitization: Absent; the skill relies entirely on the underlying agent's logic to ensure the generated Python code is safe and does not perform malicious actions.

visual-design