mage-character-sheet

Pass

Audited by Gen Agent Trust Hub on Nov 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by interpolating untrusted user data into tool prompts.
  • Ingestion points: The <details> field in SKILL.md accepts free-form text from the user for character modifications which is directly appended to the tool prompt.
  • Boundary markers: The prompt template lacks delimiters or instructions intended to prevent the agent or the downstream model from obeying instructions embedded within the user's input string.
  • Capability inventory: The skill leverages capabilities including generate (image generation), create_character (creation of persistent character profiles), and create_upload (file handling) via the Mage MCP connector.
  • Sanitization: The instructions do not implement any sanitization, filtering, or validation steps for the user-supplied 'details' before they are passed to the generation tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 13, 2026, 03:49 PM