The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill interacts with and processes data from external, untrusted sources including public websites, Instagram, and Facebook Marketplace.
Ingestion points: Data is ingested through browser.spawn_task, browser.open, instagram-cli, and facebook-cli when fetching product details, reviews, and social media post metadata.
Boundary markers: The instructions do not specify the use of delimiters or clear separation between system instructions and untrusted data fetched from the web.
Capability inventory: The skill has high-impact capabilities including the ability to navigate the web via an agentic browser (browser.spawn_task) and initiate financial transactions via shopify-ucp-cli.
Sanitization: There are no explicit instructions for the sanitization or validation of text content retrieved from external sources before it is interpreted by the agent.

shopping