nemo-automodel-model-onboarding
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists exclusively of Markdown documentation, configuration templates, and evaluation data. It does not include any executable scripts, binaries, or automated code execution patterns.
- [EXTERNAL_DOWNLOADS]: The instructions involve fetching model metadata (config.json) and model weights from the Hugging Face Hub. Hugging Face is a well-known and trusted service in the AI community, and these operations are standard for the stated purpose of model onboarding.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from Hugging Face into the agent's context. While this presents an attack surface for indirect prompt injection, the risk is negligible within the intended developer-guided workflow, and the skill includes validation phases (Phase 4 and Phase 6) that mitigate accidental or malicious configuration errors through numerical parity checks.
Audit Metadata