ml4t-stress-test
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides educational content and Python code templates for portfolio stress testing. No malicious patterns, data exfiltration, or unauthorized command execution were detected. The skill uses standard data processing libraries and defines static crisis scenarios for risk calculation.
- [INDIRECT_PROMPT_INJECTION]: The skill defines patterns to target local portfolio and risk-related files for analysis. While this establishes an ingestion surface for external project data, the skill does not implement any dangerous capabilities (such as network exfiltration, arbitrary command execution, or file system modifications) that could be exploited by malicious content within those files. All logic is contained within the local processing context.
- Ingestion points: Targets local project files matching
**/*portfolio*.py,**/*position*.py,**/*risk*.py,**/*optim*.py,**/*exposure*.py,**/*kill*.py, and**/*stress*.py. - Boundary markers: None provided as the skill is primarily an educational template.
- Capability inventory: None. The skill does not perform network operations, subprocess calls, or file writes.
- Sanitization: Not applicable as the skill provides code patterns for the user/agent to follow rather than executing untrusted data.
Audit Metadata