update-model
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: Executes
curlto perform API operations andpython3for JSON data processing. These actions are aligned with the skill's management purpose and occur locally. - [INDIRECT_PROMPT_INJECTION]: Ingests model metadata from the LiteLLM API, creating a vulnerability surface where a compromised or malicious API endpoint could attempt to inject instructions into the agent's context.
- [DYNAMIC_EXECUTION]: Uses
python3 -cto execute a static, hardcoded script for JSON transformation. While this constitutes dynamic execution, the script is transparent and limited to data processing.
Audit Metadata