amz-supplier-sample-evaluation

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow includes “Write supplier message” and “Read inserts for typos” / “Photograph any missing labels,” which implies ingesting supplier-provided sample documentation/inserts and supplier-supplied text (outsider-authored free-form content) into the agent’s LLM context for generating the message and evaluation.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I found a WhatsApp group invite link: https://chat.whatsapp.com/ILX65p1yWcaIG3c9WGHpTY. The trailing token ("ILX65p1yWcaIG3c9WGHpTY") is a high-entropy literal that grants access to the chat and is not a documentation placeholder or example, so it qualifies as a secret by the definition ("provides access to a service"). No API keys, private keys, or other high-entropy credentials were present elsewhere in the document.