prompt-engineering

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the infsh CLI tool via Bash commands to demonstrate model interaction and prompt execution. This is the intended behavior for an AI inference guide and is appropriately scoped in the frontmatter configuration.\n- [EXTERNAL_DOWNLOADS]: The skill provides links to the vendor's official GitHub repository (inference-sh/skills) for CLI installation instructions and related project-specific skills. These are trusted vendor resources.\n- [DATA_EXFILTRATION]: Execution of prompts through the provided commands sends input data to the inference.sh platform. This data transfer is the intended operational requirement for the external model inference functionality described in the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 12:27 PM