monte-carlo-simulation
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected; the skill functions as a standard statistical utility.
- [COMMAND_EXECUTION]: The skill describes simulations that may involve code execution for mathematical sampling but explicitly mandates user permission and specifies failure behaviors for missing tools like
mmdc. - [PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection through its ingestion of user-defined variables and formulas.
- Ingestion points: Processes user-supplied distributions and mathematical formulas for Monte Carlo simulations in Phases 2 and 3.
- Boundary markers: Absent; the instructions do not define delimiters for user data.
- Capability inventory: Includes file system writing for reports and diagrams, plus optional tool-based code execution for simulation logic.
- Sanitization: Absent; the skill relies on the agent's logic for parsing formulas without explicit sanitization steps defined.
Audit Metadata