Skills
Navigate
skills/amd/skills/hyperloom-workload-optimizer/Gen Agent Trust Hub

hyperloom-workload-optimizer

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the hyperloom-inference-optimizer package from the standard Python Package Index (PyPI) to prepare the optimization environment.
  • [COMMAND_EXECUTION]: The skill uses pip for package installation and executes the hyperloom-setup command to configure workspace environment variables and credentials.
  • [DYNAMIC_EXECUTION]: The installation process adds new skill definitions (hyperloom-setup and various preset run skills) to the agent's local skill directory, which are then dynamically loaded and invoked to continue the workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill monitors external optimization logs and state.json files during long-running benchmarks. To mitigate risks, the instructions include mandatory requirements to never print API keys, tokens, or custom headers in status updates, and require a human-confirmed plan before launching background processes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 05:06 PM