The Agent Skills Directory

[COMMAND_EXECUTION]: The skill manifest (SKILL.md) instructs the agent to execute a bundled Python script, scripts/cache_docs.py, to retrieve content for specific documentation topics.
[EXTERNAL_DOWNLOADS]: The cache_docs.py script fetches HTML and images from https://jeesite.com. As this is the official domain for the product being documented by the skill author, these downloads are considered legitimate.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: External HTML content is retrieved from jeesite.com and processed into Markdown for the agent (documented in all SKILL.md files under references/).
Boundary markers: The instructions define how the agent should utilize the documents but do not implement strict prompt delimiters to separate external content from instructions.
Capability inventory: The skill includes script execution, local file writing (for caching), and network access to fetch documentation.
Sanitization: The processing script includes a clean_html function to strip navigation and UI elements and employs the html2text library to ensure clean Markdown output.

jeesite