The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes a code implementation in references/performance-optimization.md that demonstrates how to dynamically load and execute an external JavaScript file from https://third-party.com/widget.js. Automated security scanners have flagged this specific URL as being associated with botnet infrastructure.
[METADATA_POISONING]: The SKILL.md file contains a documentation link targeting https://jeffallan.github.io/claude-skills/skills/platform/shopify-expert/. This URL is currently present on security blacklists, creating a risk for users who follow the project's own documentation links.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a Shopify Expert that processes data from external APIs and local theme files, creating a surface for indirect prompt injection.
Ingestion points: The skill reads local .liquid templates and fetches data via the Shopify Storefront and Admin APIs (described in references/storefront-api.md and references/app-development.md).
Boundary markers: There are no instructions to the agent to treat data from these sources as untrusted or to ignore embedded instructions within processed files.
Capability inventory: The skill utilizes the shopify CLI to perform high-privilege actions like theme push and app deploy, and executes GraphQL mutations (productCreate, metafieldsSet) to modify store data.
Sanitization: The instructions do not define sanitization or validation logic for data ingested from Shopify APIs or Liquid templates before they are used in further context or code generation.

shopify-expert