kermt-setup
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local helper script (
agent/scripts/kermt_container.sh) to perform environment checks, build container images, and run a GPU smoke test. These are standard operations for development environment automation. - [SAFE]: Analysis of the skill instructions and evaluation metadata reveals no evidence of prompt injection, data exfiltration, or obfuscation. The skill explicitly forbids unauthorized network operations (pulling/pushing images) and modifications to the host's Docker configuration or system privileges.
Audit Metadata