israeli-chatbot-analytics
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill refers to Hebrew language models (DictaBERT, AlephBERT) hosted on HuggingFace by reputable academic and research organizations (Dicta
- The Israel Center for Text Analysis, ONLP Lab). These are standard resources for Hebrew NLP tasks.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user conversation logs, which contain untrusted data. This is an inherent part of the skill's purpose. The instructions provide clear guidance on mitigating data exposure risks through pseudonymization, PII masking, and transcript retention limits.
- [COMMAND_EXECUTION]: A Python script (scripts/conversation-analyzer.py) is included for processing exported conversation logs. The script performs statistical calculations and regex-based language detection locally, with no network activity or unsafe dynamic execution patterns.
Audit Metadata