The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches HTML content from user-specified URLs for the purpose of structure analysis and data extraction. It also references technical documentation from well-known and trusted sources including 'docs.zyte.com', 'docs.scrapy.org', and 'readthedocs.io'.
[COMMAND_EXECUTION]: The skill uses 'uv run' and 'uvx' to manage virtual environments, install dependencies, and execute scraping scripts. It specifically orchestrates the execution of Scrapy spiders using 'uv run scrapy crawl'.
[REMOTE_CODE_EXECUTION]: The skill generates Python code (page objects and spiders) dynamically based on AI analysis of target websites. This generated code is executed locally to perform the scraping tasks. There is a risk that the logic generated by the AI could be influenced by malicious content on the target website to include unintended or harmful operations.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted HTML from external websites which is then processed by an AI agent to define data schemas and generate code.
Ingestion points: Target URLs provided as positional arguments to the skill in 'SKILL.md'.
Boundary markers: None explicitly defined in the instructions to separate untrusted web content from agent instructions during analysis.
Capability inventory: Access to shell execution ('Bash'), file writing ('Write'), and environment management ('uv').
Sanitization: The documentation does not specify any filtering or sanitization of the HTML content before it is processed by the AI for code generation.

scrape