engineering-ml-features
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches pre-trained machine learning models from Hugging Face's official model hub for text embedding tasks.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external datasets (text, categories, and numeric features), which creates a surface for indirect prompt injection if the processed data contains malicious instructions.
- Ingestion points: Dataframes (
df) and text collections (texts) processed acrossSKILL.mdand all reference files. - Boundary markers: Not present within the static code templates provided.
- Capability inventory: Data transformation, numeric scaling, and text vectorization. The skill does not perform unrestricted shell execution or network exfiltration to untrusted domains.
- Sanitization: Standard text preprocessing examples (lowercasing, punctuation removal) are provided in
references/text-features.mdto clean raw input data.
Audit Metadata