The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled shell script extract-and-generate.sh which coordinates the extraction process by invoking a Python script and the mmdc (Mermaid CLI) binary.
[COMMAND_EXECUTION]: The skill uses system utilities including mkdir, python3, stat, file, and ImageMagick's identify for file management and output validation.
[EXTERNAL_DOWNLOADS]: Setup instructions guide the user to install well-known software including Node.js, @mermaid-js/mermaid-cli, and Google Chrome from official repositories to satisfy dependencies.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted user-provided Markdown files to extract and render diagrams.
Ingestion points: The extract-and-generate.sh script and extract_diagrams.py read and process input Markdown files.
Boundary markers: None; diagrams are extracted based on standard Markdown triple-backtick markers.
Capability inventory: The skill can execute shell commands, run Python scripts, and invoke a headless browser via mmdc to process content.
Sanitization: There is no sanitization of the Mermaid code content before it is passed to the rendering engine.
[SAFE]: The Puppeteer configuration includes the --no-sandbox flag, which is a common requirement for running headless Chrome in restricted environments like WSL2 or CI/CD pipelines. While this reduces process isolation, it is consistent with the skill's intended use case and the requirements of the underlying tools.

mermaid-tools