The Agent Skills Directory

Prompt Injection (LOW): The skill is vulnerable to indirect prompt injection because it ingests untrusted data from multiple sources.
Ingestion points: As stated in SKILL.md (Step 1), data is gathered from direct conversation, user-provided CSV files, and website scraping.
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions for the data being processed.
Capability inventory: The skill utilizes curl for GraphQL mutations to create/update products and collections, and it generates CSV files (documented in SKILL.md and references/graphql-mutations.md).
Sanitization: Absent. There is no mention of validating or escaping data before it is interpolated into API requests or file content.
Command Execution (LOW): The skill uses curl commands in shell blocks to interact with the Shopify Admin API. These commands incorporate variables like {store} and {token} sourced from configuration files.
Data Exfiltration (LOW): The skill initiates network requests to Shopify API endpoints (e.g., https://{store}/admin/api/2025-01/graphql.json). While these requests are essential for the skill's functionality, they target domains that are not part of the standard trusted whitelist.

shopify-products