google-workspace

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

• Third-party content exposure detected (high risk: 0.85). The skill uses Google API calls to read Gmail message contents (e.g., service.users().messages().get(..., format="full").execute() and returns msg["snippet"]), where the email body/snippet is authored by external senders, so that free text can be ingested into the agent’s LLM context via the runtime message-reading path.