backtest
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill retrieves and executes the backtesting toolkit via
uvxfrom the vendor's official GitHub repository (upbit-official/upbit-strategy-toolkit). This is the intended distribution method for the author's software and is a safe practice for vendor resources. - [COMMAND_EXECUTION]: Shell commands are invoked through a dedicated wrapper script to perform backtests. The skill's instructions provide strict constraints on handling user inputs and date ranges, ensuring commands are executed within a predictable and safe workflow.
- [DATA_EXFILTRATION]: Network operations are limited to accessing historical candle data from the Upbit API, which is essential for the skill's purpose. Local file access is appropriately scoped to project-specific strategy and report directories.
- [PROMPT_INJECTION]: The skill contains comprehensive safety instructions and mandatory notices that prevent the agent from providing investment advice, evaluating strategy performance with subjective adjectives, or attempting to facilitate live trading.
Audit Metadata