The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface where user-supplied content is injected into generated HTML files, potentially leading to Cross-Site Scripting (XSS).
Ingestion points: Chart titles, subtitles, and data series provided by the user or external data sources are passed to the building scripts in scripts/build_chart.py.
Boundary markers: None detected. The skill does not use delimiters or warnings to prevent the interpretation of embedded instructions within the processed data.
Capability inventory: The skill writes these values to the filesystem as HTML files (index.html) which are served via a local web server and can be rendered in the agent's preview window. It also uses Playwright to render these pages for screenshots.
Sanitization: There is no evidence of HTML escaping or sanitization of the title, subtitle, or replacements values before they are used in string replacement or f-string interpolation within the build_chart and build_chart_custom functions in scripts/build_chart.py.

chart