The Agent Skills Directory

[DYNAMIC_EXECUTION]: The run-lint.sh script dynamically generates a temporary HTML file (runner.html) that incorporates the target document's content and the linting logic defined in lint-rules.js. This generated script is then executed within a headless Chrome environment using the --dump-dom flag to retrieve linting results.
[COMMAND_EXECUTION]: The skill includes shell scripts (scripts/run-lint.sh and scripts/render-pdf.sh) that invoke locally installed Chrome or Chromium binaries. These scripts are used to perform static analysis and PDF generation by executing the browser as a subprocess.
[EXTERNAL_DOWNLOADS]: The generated documentation artifacts reference external assets from well-known services, including Google Fonts for typography, MathJax (via JSDelivr) for mathematical typesetting, and Highlight.js (via CDNJS) for syntax highlighting.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process and edit HTML documentation, which creates a potential surface where the agent might ingest instructions embedded within the processed content.
Ingestion points: The skill reads existing HTML files and processes technical descriptions provided in the agent's context to generate documentation.
Boundary markers: There are no specific boundary markers or instructions defined to prevent the agent from following prompts embedded in the HTML content being processed.
Capability inventory: The skill has the capability to execute shell scripts (run-lint.sh, render-pdf.sh), perform filesystem writes for documentation and temporary files, and invoke a web browser (Chrome).
Sanitization: The skill uses Base64 encoding to safely pass document content into the browser-based linter, which prevents command injection into the shell call, although the linter parses the resulting HTML using standard DOM APIs.

html