design-brief

Warn

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • Indirect Prompt Injection (MEDIUM): The skill ingests untrusted external data and writes it to the local filesystem.
  • Ingestion points: User feature descriptions and external Figma design data (SKILL.md).
  • Boundary markers: Absent; no delimiters are used to separate untrusted content from system instructions.
  • Capability inventory: File-writing capability to create markdown files in the project directory (SKILL.md).
  • Sanitization: Absent; no validation for the '[name]' variable used in filenames or for the content of the brief.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 12:05 AM