The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill documentation includes a customMapFunction input parameter. This allows for the submission of a JavaScript function as a string to transform the scraper's output. Although this code is executed on the external service's infrastructure, it represents a dynamic execution vector where logic is defined at runtime through input parameters.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and return metadata from YouTube, which is an untrusted external source.
Ingestion points: Data is pulled from YouTube channel descriptions, names, and keywords based on user-provided URLs or handles.
Boundary markers: The instructions do not specify any delimiters or safety markers to wrap the retrieved content.
Capability inventory: The skill facilitates network requests to the Apify API and returns raw data for downstream processing by an agent.
Sanitization: There is no evidence of sanitization or filtering of the retrieved YouTube metadata before it is returned to the agent context.
[COMMAND_EXECUTION]: The documentation provides examples of shell commands using node and curl to trigger the scraping actor. These examples include the use of shell environment variables (e.g., $APIFY_TOKEN) for authentication and specific command-line arguments to format and save output.

scraping-youtube-channel-info