The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches HTML content and associated assets from arbitrary remote URLs provided at runtime to perform its primary scraping functions.
[COMMAND_EXECUTION]: Documentation provides instructions for installing dependencies and executing the html-get utility via npx. It also mentions the potential use of mutool for PDF-to-HTML conversion.
[REMOTE_CODE_EXECUTION]: The skill utilizes Puppeteer and Browserless to execute JavaScript within a headless browser. This is the intended mechanism for rendering JavaScript-heavy web pages and does not represent an unauthorized execution vulnerability.
[PROMPT_INJECTION]: As the skill ingests untrusted data from external websites, it presents a surface for indirect prompt injection.
Ingestion points: Arbitrary external URLs are processed via the getHTML function in SKILL.md.
Boundary markers: No specific delimiters or "ignore instructions" markers are defined for the fetched content in the provided documentation.
Capability inventory: The skill can perform network requests, render JavaScript, and rewrite HTML structure.
Sanitization: While the skill normalizes URLs and meta tags, there is no mention of sanitizing the HTML body to prevent the agent from following embedded instructions.

html-get