roofline-runtime-ru
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides technical documentation for managing inference engines (vLLM, SGLang) using a local task runner (Taskfile).
- [EXTERNAL_DOWNLOADS]: The skill references official documentation for vLLM, SGLang, and technical roadmaps from roadmap.sh. These are trusted resources for developers and system administrators.
- [COMMAND_EXECUTION]: The skill instructs the agent to use
taskcommands (e.g.,task serve,task gpu) which are expected to be defined in a localTaskfile.yml. This is standard practice for local infrastructure management and deployment automation. - [CREDENTIALS_UNSAFE]: The documentation mentions the
--api-keyflag and the use ofinference.envfiles for configuration. These are handled as placeholders or environment variables in accordance with standard best practices for secret management, with no hardcoded credentials detected.
Audit Metadata