co-op-game-intro-generator

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data, including player names, game titles, and custom visual style descriptions, which are interpolated into prompt templates in the references directory. This creates a surface for indirect prompt injection.
  • Ingestion points: User-provided player names, game title, and visual style strings collected in SKILL.md (Steps 1 and 2).
  • Boundary markers: The skill utilizes structured template files (e.g., references/h3-confirmation-image-template.md) which act as delimiters between the fixed UI framework instructions and dynamic user-selected style data.
  • Capability inventory: The skill's capabilities are limited to generating images and videos via platform-native generation tools (MiniMax H3).
  • Sanitization: No explicit input validation or escaping logic is described in the instructions, though the templates include extensive negative constraints to manage generation quality and safety.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 02:58 AM