hedge-risk-reward
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute and potentially modify local Python scripts (
backtest_<slug>.py) located within the.cache/hedge-backtesting/directory. This is intended to facilitate the extraction of trade lists and equity curves necessary for analysis.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing external data artifacts such asrun.log, trade-level CSVs, and stdout tables. Maliciously crafted data in these sources could theoretically influence the agent's analysis or output summary.
Audit Metadata