bagisto-api-admin

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill reads outsider-authored free text at runtime via Bagisto Admin API listing/detail endpoints such as the Reviews moderation queue (originating from storefront authors) and order comments, which the required workflow can fetch and display to the LLM/UI without selecting a specific item.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch and consult the Bagisto API docs at runtime—calling out https://api-docs.bagisto.com and its /llms.txt index as "the source of truth" and to "fetch llms.txt" / open docs pages before writing calls—so external content would be fetched during runtime and directly control prompt/implementation choices.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly targets the Bagisto Admin API and documents admin actions that perform financial operations: placing orders (Create-Order), creating invoices, processing refunds, and managing transactions. These are specific, built-in API actions (REST/GraphQL mutations) intended to execute commerce-related money flows from an admin interface, not generic tooling—so it grants direct financial execution capability.