NYC

nemo-curator

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill utilizes standard NVIDIA RAPIDS ecosystem libraries such as nemo-curator, cudf, and dask. Installation is handled through reputable package managers and no suspicious execution patterns or untrusted sources were detected.
  • Indirect Prompt Injection (SAFE): While the curation pipeline processes external data (e.g., Common Crawl), the skill specifically includes safety modules for PII redaction and NSFW detection to mitigate risks. Evidence: 1. Ingestion points: DocumentDataset.read_parquet in SKILL.md. 2. Boundary markers: None. 3. Capability inventory: Data transformation and local file writing. 4. Sanitization: PIIRedactor and NSFWClassifier modules are natively integrated.
  • Data Exposure & Exfiltration (SAFE): No evidence of hardcoded credentials, sensitive file path access, or unauthorized network exfiltration.
  • Obfuscation (SAFE): No encoded content, hidden characters, or obfuscated logic were found in any files.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:58 PM