gws
Pass
Audited by Gen Agent Trust Hub on Mar 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the
@anthropic-ai/gwsCLI tool using npm to enable communication with Google APIs. - [COMMAND_EXECUTION]: The skill frequently executes the
gwscommand to perform actions like searching emails, reading messages, and managing calendar events. - [PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by processing external data such as email content and calendar descriptions.
- Ingestion points: Gmail message bodies retrieved via
gws gmail users messages getand calendar event details viagws calendar events list. - Boundary markers: No specific delimiters or "ignore instructions" warnings are provided in the prompt templates to separate untrusted data from system instructions.
- Capability inventory: The skill possesses the ability to send emails (
gws gmail +send), reply to messages, and create/delete calendar events. - Sanitization: No explicit sanitization or content validation is performed on the retrieved data before it enters the agent's context.
Audit Metadata