The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external command-line tools to render diagrams.
scripts/render_mermaid.py executes the mmdc (Mermaid CLI) tool via subprocess.run using input files generated from agent content.
scripts/render_mindmap.py executes the markmap CLI tool via subprocess.run to convert Markdown to SVG/PNG.
[EXTERNAL_DOWNLOADS]: Generated HTML files for ECharts, DrawIO, and GeoGebra reference scripts and resources from external, well-known services.
ECharts templates use cdn.jsdelivr.net for the core library.
GeoGebra templates load the deployment API from www.geogebra.org.
DrawIO templates embed an iframe pointing to embed.diagrams.net.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8).
Ingestion points: User-provided text is processed by the agent to generate code for Mermaid, ECharts, Mindmap, DrawIO, or GeoGebra.
Boundary markers: There are no explicit boundary markers or 'ignore embedded instructions' warnings in the generated configuration files.
Capability inventory: The skill possesses file-writing capabilities and the ability to execute subprocesses (mmdc, markmap) on the generated content.
Sanitization: While scripts/render_echarts.py and scripts/render_drawio.py use JSON serialization to safely escape data for HTML/JavaScript generation, the underlying diagram code (like Mermaid or GeoGebra commands) is passed directly to third-party rendering engines which may have their own parsing vulnerabilities.
[NO_CODE]: Several scripts serve as templates for dynamic HTML generation.
scripts/render_geogebra.py and scripts/render_drawio.py generate web pages that execute logic in the user's browser context based on agent-generated parameters.

multi-chart-draw