together-gpu-clusters
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions in
references/tcloud-cli.mdto download the standalonetcloudCLI binary from an Amazon S3 bucket (tcloud-cli-downloads.s3.us-west-2.amazonaws.com). This is a documented official distribution method for Together AI's standalone tooling. - [COMMAND_EXECUTION]: The skill includes scripts and documentation for executing shell commands and API calls to provision, scale, and delete GPU clusters. These capabilities are central to the skill's purpose of managing high-performance computing infrastructure.
- [CREDENTIALS_UNSAFE]: The documentation references the use of the
TOGETHER_API_KEYenvironment variable and provides instructions for retrieving Kubernetes admin tokens and managing SSH keys for cluster access. These represent standard administrative procedures for cloud infrastructure and do not include hardcoded secrets. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process status outputs and health check data from the provisioned clusters (e.g., node health, GPU diagnostics). While this introduces a theoretical ingestion point for untrusted data, the risk is mitigated by the administrative context and the use of the official SDK for structured interactions.
Audit Metadata