canvas-design

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions allow the agent to download fonts as needed for artwork creation. The accompanying license files in the canvas-fonts/ directory refer to well-known and reputable repositories such as Google Fonts, IBM, and Vercel Labs. These external references are standard for design tasks and align with the skill's primary purpose.
  • [COMMAND_EXECUTION]: Generating visual assets and retrieving typography resources involves routine shell and file operations. These actions are documented and essential for the skill's intended functionality of creating high-quality PDF and PNG artifacts.
  • [PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by ingesting user conceptual input to generate a design philosophy.
  • Ingestion points: User-provided inputs in SKILL.md serve as the foundation for the 'Design Philosophy Creation' step.
  • Boundary markers: Explicit delimiters for separating user input from the rest of the philosophy prompt are absent.
  • Capability inventory: The skill has the ability to create and write local files (.png, .pdf, .md).
  • Sanitization: No specific validation or filtering of user input is described in the instructions. This input ingestion is necessary for the creative workflow and does not represent a malicious attempt to bypass safety controls.
  • [SAFE]: All components of the skill, including the license information and design instructions, are consistent with its stated goal of providing a creative design workspace.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 12:19 PM
Security Audit — agent-trust-hub — canvas-design