The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The file references/performance-optimization.md contains a script reference to https://third-party.com/widget.js. This URL was flagged by automated scanners as being associated with botnet activity.
[REMOTE_CODE_EXECUTION]: The guidelines in references/performance-optimization.md provide a JavaScript pattern to dynamically inject and execute external code from the flagged domain third-party.com. This represents a critical remote code execution vector.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Untrusted data ingested via the Shopify Storefront API and Admin API (e.g., product titles, descriptions, and metafields) as referenced in SKILL.md and references/storefront-api.md.
Boundary markers: Absent. There are no instructions to use delimiters or ignore instructions embedded within the API response data.
Capability inventory: Execution of Shopify CLI commands (shopify theme, shopify app), network fetch operations, and dynamic script loading.
Sanitization: Absent. The skill lacks explicit sanitization or validation protocols to prevent the agent from following instructions embedded in the third-party storefront data.
[COMMAND_EXECUTION]: The workflow relies heavily on the Shopify CLI tool suite, which has significant permissions to modify the local environment and interact with network services. While standard for Shopify development, these capabilities are highly exploitable in the presence of malicious instructions.

shopify-expert