management-sprint-retro

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 ingests team-provided retro notes (silent writing / sticky notes / board entries) into the agent’s LLM context; these notes are outsider-authored free text if they come from people other than the operating user (e.g., teammates), creating an indirect prompt-injection risk via untrusted human text.