unsloth-tokenizer

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill examples demonstrate fetching model and tokenizer configurations from Hugging Face Hub repositories (e.g., 'meta-llama/Llama-3.2-1B', 'mistralai/Mistral-7B-v0.1', 'google/gemma-2b'). Hugging Face is a well-known service for hosting AI artifacts, and these operations are standard for the described use case.- [DATA_EXFILTRATION]: The skill includes code for reading local files such as 'test_corpus.txt', 'domain_corpus.txt', and local JSON tokenizer files. These operations are restricted to data ingestion for local analysis and benchmarking, with no signs of unauthorized external transmission.- [REMOTE_CODE_EXECUTION]: The skill relies on well-known and reputable machine learning libraries including 'unsloth' and 'transformers'. These packages are industry standards for LLM fine-tuning and tokenization analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 11:11 AM
Security Audit — agent-trust-hub — unsloth-tokenizer