Skills
Navigate
skills/vesviet/agent-skills/generate-game-assets-pipeline/Gen Agent Trust Hub

generate-game-assets-pipeline

Pass

Audited by Gen Agent Trust Hub on Oct 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting data from external AI-generation services and asset archives, creating a potential indirect prompt injection surface.
  • Ingestion points: AI-generation service outputs, prompt metadata, and downloaded texture/model archives (referenced in SKILL.md).
  • Boundary markers: The skill includes explicit instructions (ASI01) to treat external data as non-executable content and never as instructions.
  • Capability inventory: The skill utilizes execute_command for running asset processing tools such as ImageMagick and Pillow (referenced in SKILL.md).
  • Sanitization: The instructions recommend path validation, tool version pinning, and checksum verification (ASI04, ASI05) to ensure integrity.
  • [COMMAND_EXECUTION]: The skill workflow requires the agent to execute shell commands for processing and validating game assets.
  • Evidence: References to using Aseprite CLI, ImageMagick, and Python scripts via execute_command for batch processing and resolution validation in SKILL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 31, 2026, 11:20 AM