connector-googlemail
Fail
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill specifies dependencies on external Motoko packages
googlemail-client,google-oauth, andcaffeineai-authorizationfrom the mops.one registry. These are standard libraries for the Motoko development ecosystem. - [COMMAND_EXECUTION]: The documentation provides shell commands (
mops add) for developers to install the required package dependencies. - [INDIRECT_PROMPT_INJECTION]: The skill implements a messaging interface that processes user-supplied text for email content, which is a potential surface for indirect prompt injection if the inputs are not validated or delimited by the calling agent.
- Ingestion points: Email
subjectandbodyfields in thesendEmailmixin. - Boundary markers: Not explicitly defined in the provided code logic.
- Capability inventory: Includes network access to Gmail REST API endpoints for sending messages.
- Sanitization: Relies on provided libraries for data serialization and encoding for the Gmail API.
Recommendations
- Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata