stride-analysis-patterns

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): No instructions to bypass safety filters or override agent constraints were found.
  • [Data Exposure] (SAFE): No sensitive file paths, environment variable access, or hardcoded credentials detected.
  • [Network Activity] (SAFE): The skill does not perform any network operations or data exfiltration.
  • [Remote Code Execution] (SAFE): No external scripts, packages, or remote execution patterns are present.
  • [Indirect Prompt Injection] (LOW): The skill is designed to process user-provided architectural data. While this is an ingestion surface, it lacks the capabilities (exec/eval/network) to weaponize such an injection. Usage of boundary markers for external data is recommended.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:12 PM