theme-builder
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes data from local theme files (
blocks/*.liquid,sections/*.liquid), which is an inherent part of its functionality. This is a surface for indirect prompt injection if those project files contain malicious instructions, though the risk is negligible for the intended use case.\n - Ingestion points:
SKILL.md(Workflow Step 1 and 3) instructs reading of Liquid schema blocks.\n - Boundary markers: The agent is not explicitly told to ignore instructions inside schemas, but focuses on parsing JSON parameters.\n
- Capability inventory: Reading workspace files and writing Shopify JSON templates.\n
- Sanitization: No specific content sanitization is mentioned, as the tool relies on standard Shopify schema structure.
Audit Metadata