google-workspace

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

• Third-party content exposure detected (medium risk: 0.30). The Google Workspace skill integrates access to Gmail, Google Chat, and Google Drive, which can ingest outsider-authored free text (emails, chat messages, and document content) during runtime execution.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill references an external cloud function URL at https://google-workspace-extension.geminicli.com for token refresh and OAuth redirection without first-party cryptographic verification, constituting a weak/unknown external operational dependency.