pencil-design

Warn

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill instructs the agent to author and execute JavaScript snippets via the batch_design tool. Additionally, it directs the creation of .js script nodes and .glsl shaders that are executed within the Pencil runtime environment, representing a significant dynamic code generation surface.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from .pen design files (JSON format) and design-system/ markdown documentation. It lacks specific boundary markers or instructions to disregard potential commands embedded within this data. This exposure, combined with the capability to execute code via batch_design, creates a surface for indirect prompt injection where malicious design files could influence agent behavior.
  • [EXTERNAL_DOWNLOADS]: The documentation references the @pencil.dev/cli tool and various established design systems (Tailwind, Radix Colors, IBM Carbon, etc.). These references are to the tool vendor's own packages and well-known, trusted industry services.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 14, 2026, 10:43 PM