aws-iot-greengrass

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches the AWS Greengrass Lite runtime source code and the Component SDK from the official AWS GitHub organization repositories.\n- [COMMAND_EXECUTION]: The instructions include numerous shell commands for managing Podman/Docker containers, building C++ components with CMake, and performing resource provisioning via the AWS CLI.\n- [PRIVILEGE_ESCALATION]: The setup guide for the Greengrass Nucleus runtime specifies running the container as root and using the --privileged flag, which is a documented requirement for the runtime's core operations.\n- [DATA_EXFILTRATION]: The skill includes functional examples of components that monitor local directories to upload files to Amazon S3 and publish telemetry to AWS IoT Core as part of intended IoT functionality.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies surfaces for indirect prompt injection by defining components that ingest and process messages from external IoT topics or local inter-process communication channels.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 12:54 PM