awq-quantization
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.90). This skill explicitly loads models/tokenizers from the public HuggingFace model hub (e.g., AutoAWQForCausalLM.from_quantized("TheBloke/..."), AutoTokenizer.from_pretrained(...)), accepts public calibration datasets by name (e.g., "pileval", "wikitext") and links to GitHub, so it ingests untrusted, user-published third-party content that the agent will read/interpret.
Audit Metadata