make-game

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from project-level markdown documents and user input to drive its logic and generate code. Ingestion points: Reads from project documentation in the docs/ directory, including gameplan.md, tech.md, milestones/, and backlog.md. Boundary markers: The skill employs a strict multi-phase document-driven workflow which serves as a logical context boundary. Capability inventory: The skill can perform file system writes and execute shell commands for scaffolding and project management. Sanitization: Verification is enforced through mandatory live-iterate loops and automated browser-based testing.
  • [COMMAND_EXECUTION]: The agent is instructed to execute various shell commands for project scaffolding and dependency management. Evidence: The skill references commands such as npm create @phaserjs/game@latest, npm install three, and official CLI tools for engines like Unity and Godot. These actions are standard for the development and setup of software projects.
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for downloading and installing additional specialized skills from the opusgamelabs organization. Evidence: Usage of the command npx skills add opusgamelabs/game-creator --skill is recommended to extend functionality. These downloads are part of the skill's own tooling ecosystem and originate from the author's organization.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 12:09 PM