shop-product-launch
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input that is passed to shop management tools, creating a surface for embedded instructions to influence agent behavior.
- Ingestion points: The 'Input contract' in
SKILL.mdidentifies Name, short description, long description, and editorial paragraphs as untrusted data sourced from the user. - Boundary markers: The instructions lack specific delimiters or warnings to the agent to treat this content as data rather than instructions.
- Capability inventory: The workflow involves multiple write operations across ALL scripts in
SKILL.md, includingshop_create_product,shop_create_option,shop_create_variant, andshop_create_paragraph. - Sanitization: There is no requirement or implementation of content sanitization or escaping for the textual fields before they are processed by the tools.
Audit Metadata