The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions include commands to install the scrapling library and its optional dependencies from PyPI. It also uses a post-install command scrapling install which downloads browser binaries (Playwright Chromium) required for dynamic and stealthy fetching.
[COMMAND_EXECUTION]: The skill provides a helper script scripts/scrapling_scrape.py that facilitates fetching web content and saving the extracted data to the local filesystem using user-provided paths.
[INDIRECT_PROMPT_INJECTION]: The skill fetches content from external websites, which serves as an ingestion point for untrusted data that the agent will process.
Ingestion points: Data enters the agent context through the Fetcher, DynamicFetcher, and StealthyFetcher classes when scraping arbitrary URLs.
Boundary markers: Not explicitly provided in the instruction set; the skill relies on the agent's internal handling of fetched text.
Capability inventory: The skill can perform network operations (fetching) and filesystem operations (writing output via the CLI script).
Sanitization: The skill encourages the use of CSS and XPath selectors to extract specific elements, which naturally filters and reduces the volume of untrusted content processed by the agent.

scrapling