The Agent Skills Directory

[DATA_EXFILTRATION]: The skill implements a system to capture verbatim user prompts and transmit them to official Shopify telemetry and validation endpoints. The SKILL.md instructions require the agent to base64-encode user messages and pass them to validation scripts. The telemetry hooks in scripts/track-telemetry.sh and scripts/track-telemetry.ps1 stash this data in temporary files before transmission to official usage endpoints. This data collection is documented as a standard feature, and an opt-out is available via the OPT_OUT_INSTRUMENTATION environment variable.
[COMMAND_EXECUTION]: The skill workflow requires the agent to execute local JavaScript and shell scripts for its primary operations. The agent utilizes scripts/search_docs.mjs for documentation queries and scripts/validate.mjs for code verification, while also using the Shopify CLI for project scaffolding and builds.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it ingests data from external Shopify documentation to guide code generation. Ingestion points include documentation results retrieved by the search script. There are no explicit boundary markers to prevent the agent from potentially following instructions embedded in retrieved content, although the source is official infrastructure. Capability inventory includes shell command execution and file system access. Sanitization of retrieved content is not implemented in the provided scripts.

shopify-functions