The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Excel and CSV files, which serves as an ingestion point for untrusted data. A malicious file could attempt to manipulate the analysis results or embedded insights generated by the agent.
Ingestion points: The scripts/analyzer.py script loads data from user-supplied file paths provided to the agent.
Boundary markers: Absent; there are no specific instructions or delimiters used to ensure the agent ignores malicious content within the analyzed files.
Capability inventory: The skill has the capability to read local data files, execute Python analysis scripts, and write generated HTML reports to the workspace.
Sanitization: Data is parsed using the pandas library. The HTML generation process uses JSON.stringify to serialize chart configurations before they are executed in the browser context.
[DYNAMIC_EXECUTION]: The assets/report_template.html file includes logic that uses eval() to initialize interactive charts based on the processed analysis data.
Evidence: eval(chartInitCode); is located in the JavaScript section of the HTML template.
Context: This is used to hydrate ECharts options that are generated dynamically based on the characteristics of the input data.
[EXTERNAL_DOWNLOADS]: The generated report fetches the ECharts library from a public content delivery network.
Evidence: <script src="https://cdn.jsdelivr.net/npm/echarts@5.5.0/dist/echarts.min.js"></script> in assets/report_template.html and referenced in references/chart_types.md.
Context: The library is fetched from cdn.jsdelivr.net, which is a well-known service for hosting public open-source assets.

data-analysis-report-generator