The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from YouTube, including video titles and descriptions, to identify trends and title formulas. This external content is ingested without sanitization or the use of boundary markers, which could allow malicious instructions embedded in video metadata to influence the agent's behavior.
Ingestion points: YouTube video titles, descriptions, and engagement metrics retrieved through the scraper.
Boundary markers: The skill does not define delimiters or specific instructions for the agent to ignore commands within the ingested text.
Capability inventory: The skill uses shell commands (node, curl) and an MCP tool, creating a path for potential exploitation if the agent follows instructions found in metadata.
Sanitization: No validation or escaping of the raw YouTube data is mentioned before it is processed by the analysis logic.
[DYNAMIC_EXECUTION]: The skill exposes a customMapFunction parameter, which allows for a JavaScript string to be passed and executed on the Apify platform to transform data results. While this is a standard feature of the underlying scraping actor, it represents a runtime code execution vector.
[COMMAND_EXECUTION]: The skill provides instructions for executing shell commands via a local script scripts/run_actor.js and curl to interact with the Apify API for data retrieval and actor management.

tracking-youtube-trending-topics-by-niche