The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references official SDKs for the Scavio API, including the scavio package on PyPI and npm, as well as the @scavio/mcp-server. These are legitimate, versioned resources provided by the skill author for interacting with the service.
[INDIRECT_PROMPT_INJECTION]: The skill processes product data, specifications, and customer reviews fetched from Wayfair marketplaces.
Ingestion points: Untrusted external data enters the agent context through the /search, /product, /category, and /deals endpoints of api.scavio.dev.
Boundary markers: The skill provides detailed instructions on how to handle the data, such as quoting AI-generated labels and distinguishing between organic and sponsored placements to maintain context integrity.
Capability inventory: The skill focus is on data retrieval and display; there are no instructions or scripts that would allow the agent to execute shell commands or write to the file system based on the ingested content.
Sanitization: Instructions explicitly command the agent to never fabricate data and to return only the values provided by the API, which serves as a prompt-level constraint against halluncination or injection-based behavior.

wayfair-api