The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a potential surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through the extract_website(url), extract_brand_kit(url), extract_brand_asset(url), and extract_photography(urls) tools, which fetch content from arbitrary user-provided or search-discovered URLs.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore embedded instructions within the fetched metadata or website content.
Capability inventory: The skill enables the agent to fetch and read external content (get_website_result, get_brand_kit_result) to inform design decisions, which could influence downstream agent actions if the source content contains malicious instructions.
Sanitization: There are no instructions for the agent to sanitize, escape, or validate the content retrieved from external sites before it is incorporated into the prompt context.

brand-search