ml4t-kill-switch
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides documentation and Python code examples for creating a trading kill switch that monitors portfolio risk metrics.
- [SAFE]: All identified logic is consistent with the stated purpose of risk management and does not include hidden commands, unauthorized data access, or malicious persistence mechanisms.
- [DATA_EXPOSURE]: The skill's code examples use standard variable placeholders for credentials (e.g.,
api_key,secret_key). It follows security best practices by explicitly instructing the user to retrieve reset codes from a secret store rather than hardcoding them in the source code. - [INDIRECT_PROMPT_INJECTION]: The skill processes external numerical data (such as P&L and drawdown) to trigger risk-mitigation logic. This represents a standard functional requirement for trading systems. Because the risk thresholds are evaluated using hard-coded Python comparisons rather than natural language processing by an LLM, the logic is not susceptible to prompt injection via data feed manipulation.
Audit Metadata