ops-inspector

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime uses user-provided query parameters to call queryLogs(action="searchLogs", ...) and queryFunctions(action="listFunctionLogs", ...) to ingest log text and error messages from the tenant’s CLS/system logs into the LLM for inspection, so an outsider can supply free text via the workflow’s search/query inputs (indirect prompt injection through stored logs or injected search terms).