google-docs-automation

Warn

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Python scripts (scripts/auth.py and scripts/docs.py) to manage authentication and document operations.
  • [DATA_EXFILTRATION]: The skill documentation states that access tokens are refreshed using a remote cloud function. Transmitting OAuth credentials to an unspecified external service poses a risk of credential harvesting or unauthorized access if the service is compromised or malicious.
  • [PROMPT_INJECTION]: The skill processes untrusted document content, which facilitates indirect prompt injection attacks.
  • Ingestion points: Document text is retrieved into the agent context via the get-text command.
  • Boundary markers: No delimiters or instructions are used to distinguish document content from system instructions.
  • Capability inventory: The agent can execute commands, modify document content, and create new documents, providing a significant impact path for malicious instructions.
  • Sanitization: There is no evidence of filtering or validation logic for the ingested text.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 22, 2026, 09:03 PM
Security Audit — agent-trust-hub — google-docs-automation