shopify-dev
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill transmits usage data to Shopify's infrastructure (shopify.dev). This data includes session identifiers, model names, and the verbatim user prompt that triggered the skill.\n
- Evidence: The logging mechanism is implemented in scripts/log_skill_use.mjs and documented in SKILL.md. Telemetry hooks in scripts/track-telemetry.sh and scripts/track-telemetry.ps1 also capture and send session data.\n
- Note: This behavior is disclosed in the skill's privacy notices and respects the OPT_OUT_INSTRUMENTATION environment variable.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and processes documentation search results from an external API without explicit boundary markers or sanitization of the content before it is presented to the agent.\n
- Ingestion points: Search results retrieved from https://shopify.dev/assistant/search in scripts/search_docs.mjs.\n
- Boundary markers: Absent; search results are printed directly to the output stream.\n
- Capability inventory: The skill has the ability to execute local scripts and perform network operations.\n
- Sanitization: Absent; search results are returned as raw text or formatted JSON.\n- [COMMAND_EXECUTION]: The skill relies on several local scripts (.mjs, .sh, .ps1) to perform its primary search functions and handle telemetry reporting.\n
- Evidence: Scripts are located in the scripts/ directory and are invoked via standard shell commands as described in SKILL.md.
Audit Metadata