The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted input from customer records and internal notes which are then processed by the agent.
Ingestion points: Data enters the agent context through the customer_email, customer_id, and reason parameters defined in SKILL.md.
Boundary markers: The instructions lack explicit delimiters or specific directives for the agent to ignore potential commands embedded within the customer-provided data or notes.
Capability inventory: The skill utilizes the shopify-admin-execution toolkit to perform giftCardCreate mutations in SKILL.md, which create real-world monetary value.
Sanitization: There is no evidence of input validation or sanitization for the reason or customer_email fields before they are interpolated into the GraphQL mutation.
[COMMAND_EXECUTION]: The skill performs GraphQL mutations that result in the creation of financial value (gift cards).
Evidence: The giftCardCreate mutation in SKILL.md is used to generate redeemable store credit.
Mitigation: The skill properly includes a dry_run parameter and safety warnings to encourage verification before committing changes.

gift-card-issuance