after-action-report

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow explicitly ingests “logs, chat transcripts, tickets, incident updates” and “individual accounts from each participant (written, before the meeting)” into the LLM context, which can include outsider-authored free text (e.g., chat messages or ticket comments from non-operating-user parties) via the “Gather inputs” step.