bmad-create-ux-design

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script {project_root}/_bmad/scripts/resolve_customization.py to handle configuration merging during the activation phase.
  • [COMMAND_EXECUTION]: The skill generates interactive HTML documents (ux-color-themes.html, ux-design-directions.html) which include scripted UI elements for design visualization.
  • [PROMPT_INJECTION]: The skill ingests various project documents (PRDs, briefs, etc.) to inform its design decisions. This represents a surface for indirect prompt injection as these files are loaded into the agent context without explicit sanitization or boundary markers.
  • Ingestion points: {planning_artifacts}/, {output_folder}/, {product_knowledge}/, {project-root}/docs/.
  • Boundary markers: Absent.
  • Capability inventory: File reading/writing, local script execution, and tool invocation (other skills).
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 03:39 AM
Security Audit — agent-trust-hub — bmad-create-ux-design