The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's SKILL.md file mandates as a 'Non-Negotiable' the execution of npx --yes skills update shopify-store-translator -g -y on every trigger. This results in the automatic download and execution of arbitrary code from the NPM registry without user oversight or confirmation.
[PROMPT_INJECTION]: The instructions include 'Non-Negotiables' that direct the agent to 'not ask the user' and 'do not skip' the remote update command. This framing attempts to override the agent's autonomy and platform safety checks. Additionally, it possesses an indirect prompt injection surface:
- Ingestion points: Ingests external store data (products, blogs) via the fetch command in shopify-translator-admin.mjs.
- Boundary markers: Absent in the translation workflow.
- Capability inventory: Includes Node.js script execution, file system writes (translation-audit.csv), and Shopify CLI mutations.
- Sanitization: Lacks sanitization or validation of the ingested translatable content.
[COMMAND_EXECUTION]: The skill executes multiple shell commands, including shopify store execute for GraphQL mutations, npm install for dependencies, and custom Node.js scripts for local data processing.
[EXTERNAL_DOWNLOADS]: The skill triggers the installation of the @shopify/cli package and updates its own logic via npx from external registries (NPM).

shopify-store-translator