The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The file references/performance-optimization.md contains a JavaScript code block intended to dynamically inject and execute an external script from https://third-party.com/widget.js. This URL is explicitly flagged as malicious (associated with botnet activity) by security scanners, posing a high risk to any application that implements this pattern.
[EXTERNAL_DOWNLOADS]: The skill recommends and provides templates for fetching remote code from untrusted third-party domains without implementing security controls such as Subresource Integrity (SRI).
[INDIRECT_PROMPT_INJECTION]: The skill presents a significant attack surface by processing unvalidated user requirements and external data while possessing powerful system capabilities. Ingestion points: SKILL.md (Workflow Step 1: Requirements analysis) and references/storefront-api.md (GraphQL query parameters constructed from external inputs). Boundary markers: None identified in the skill body. Capability inventory: SKILL.md (Shopify CLI commands for deployment), references/app-development.md (Admin API mutations), and references/checkout-customization.md (Wasm compilation via cargo wasi build). Sanitization: No validation or sanitization of ingested content or external scripts is mentioned.

shopify-expert