The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The instructions recommend installing the basalam-mcp package using uvx, which involves downloading and running executable code from a public package registry.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from an external marketplace, which is a source of untrusted data.
Ingestion points: Untrusted data enters the agent context through marketplace search results, product attributes, shop descriptions, and customer reviews fetched by tools such as bs_search, bs_product, bs_vendor, and bs_reviews.
Boundary markers: The skill does not define specific delimiters or include instructions to the agent to ignore potentially malicious directions embedded within the marketplace data.
Capability inventory: The tools utilize network operations to communicate with marketplace APIs.
Sanitization: There is no mention of sanitization, escaping, or validation of the text retrieved from the external marketplace before it is processed by the AI.

basalam