shopify-shopifyql

SUSPICIOUS. The skill's core analytics/query capability matches its stated purpose and uses same-org Shopify infrastructure, but it also mandates telemetry that sends the user's verbatim prompt and host metadata to Shopify on every use. This is proportionally broader than necessary for a ShopifyQL-writing skill, though not fundamentally malicious or routed to unrelated third parties.

No clear evidence of classic supply-chain sabotage (no obfuscated payload execution, no reverse shells, no persistence, no eval). The code appears to be a telemetry hook that parses untrusted JSON input, stashes and later transmits user prompt text (up to 2000 chars) plus session/tool metadata to a telemetry endpoint via Invoke-RestMethod in a child PowerShell process. The primary risks are privacy/data-exfiltration and endpoint integrity; since the $endpoint definition/validation is not included in the fragment, the likelihood that data could be sent to an attacker-controlled destination cannot be fully ruled out.

The code implements explicit Shopify usage telemetry and sends decoded user-prompt content, session metadata, and client metadata to Shopify-controlled or environment-selected endpoints. This creates a meaningful privacy and data-leakage concern if prompts or session identifiers are sensitive, and the configurable instrumentation URL permits local redirection of telemetry. The behavior appears intentional and related to instrumentation rather than malware. Users should verify organizational consent and configure the documented opt-out mechanisms before use.