design-system

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it ingests untrusted data from the workspace and has capabilities to generate content and export files.
  • Ingestion points: The skill is instructed to read the .agents/marketing-context.md file to establish brand identity and visual guidelines.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the context file as untrusted or to ignore embedded instructions.
  • Capability inventory: The skill uses the mcp__pencil__batch_design tool (which can generate content via AI prompts) and mcp__pencil__export_nodes to export design assets.
  • Sanitization: There is no evidence of sanitization, validation, or filtering of the content retrieved from the marketing context file before it is used in design operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 10:00 AM
Security Audit — agent-trust-hub — design-system