product-ops

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime “Work mode” operations (e.g., VoC/feedback synthesis and other record-touching sub-workflows) fetch and process Airtable records whose fields like Verbatim/Summary can contain free-form text submitted by outsiders (e.g., customer feedback, support tickets, Slack messages) and then feed that text into the agent/LLM for classification and summarization.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)