privilege-log-review

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow reviews a privilege log and “reads entries and underlying documents” to classify them, so any outsider-authored log text (e.g., opposing party’s privilege log entries or other third-party communications included in the provided log/documents) would be ingested as readable prose into the agent’s LLM context during Step 2 entry-by-entry.