NYC

data-viz-plots

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTIONNO_CODE
Full Analysis
  • [Indirect Prompt Injection] (HIGH): The skill processes untrusted data for visualization purposes. \n
  • Ingestion points: External datasets provided for plotting. \n
  • Boundary markers: Unknown/Absent due to missing implementation code. \n
  • Capability inventory: Execution of Python-based plotting code. \n
  • Sanitization: Unknown/Absent. \n- [Dynamic Execution] (HIGH): The skill generates and executes Python code using matplotlib and seaborn. Without isolation or validation, this represents a risk for arbitrary command execution. \n- [Metadata Poisoning] (HIGH): The skill metadata describes powerful functionality without providing the core logic in SKILL.md, making it impossible to verify claims of safety or intended behavior.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 03:38 AM