perps-write-ticket

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

• Third-party content exposure detected (high risk: 0.70). The required workflow is to “read the raw input” (e.g., Slack message/screenshot/one-line idea) and draft a ticket from it, so if that raw input is authored by someone other than the operating user, that outsider free text is ingested into the agent’s LLM context during intake.