The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites, which creates an ingestion surface for potential indirect prompt injection if the output is passed to an LLM. However, the skill provides specific, mandatory instructions in Section 6.0a to sanitize scraped content by stripping hidden or executable elements like script and style tags before processing.
[EXTERNAL_DOWNLOADS]: The documentation references several well-known scraping and parsing libraries, such as Playwright, Scrapling, and readability-lxml. It also provides instructions for utilizing the Microsoft Webwright framework, which is a trusted repository, for complex navigation tasks.
[COMMAND_EXECUTION]: The skill uses Playwright and standard Python scripts to automate browser interactions and local file operations. These operations are conducted using well-documented libraries and are scoped to the skill's primary function of harvesting and saving structured data.

website-scraping