The Agent Skills Directory

[SAFE]: The skill implements standard scientific visualization workflows using reputable libraries such as matplotlib, seaborn, and plotly. All scripts and instructions focus exclusively on data plotting and figure styling to meet academic publication standards.\n- [SAFE]: No indicators of malicious activity such as prompt injection, data exfiltration, or persistence mechanisms were found. The skill operates locally and does not perform network operations to unknown or suspicious domains.\n- [SAFE]: The provided scripts for figure export and style presets use standard Python APIs and do not contain dangerous patterns like eval(), exec(), or arbitrary subprocess execution. External references point to well-known academic and accessibility resources.\n- [SAFE]: Indirect Prompt Injection Risk Assessment: \n
Ingestion points: User-provided data structures (DataFrames and arrays) enter the system through plotting functions in matplotlib_examples.md.\n
Boundary markers: None are explicitly used for input data. \n
Capability inventory: The skill's capabilities are restricted to local file system writes (savefig) in figure_export.py. No network access, shell execution, or code evaluation tools are present.\n
Sanitization: None; however, the lack of high-privilege capabilities or outbound data channels mitigates the risk from malicious data input.

scientific-visualization