The Agent Skills Directory

[SAFE]: The skill provides instructional content and code templates for Magento development that follow established security best practices. Specifically, it recommends storing authentication tokens in httpOnly cookies rather than localStorage to mitigate the risk of token theft via Cross-Site Scripting (XSS). It also demonstrates proper GraphQL error handling by checking for the errors property in the response body.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from an external Magento GraphQL API into the application context, which represents an indirect prompt injection surface if the data is processed by an LLM without proper boundaries.\n
Ingestion points: Data returned by the magentoQuery function (e.g., product lists, cart items, customer details) in SKILL.md.\n
Boundary markers: Not explicitly defined in the templates, as they focus on API connectivity and data structure.\n
Capability inventory: The skill enables network operations via fetch to interact with the Magento /graphql endpoint.\n
Sanitization: The provided code snippets include validation of the GraphQL errors array and HTTP response status checks.

magento-graphql