goose-adventure-game

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its reliance on external data files for game logic and state. 1. Ingestion points: The engine reads data from scripts/game_script.json and dynamically loads save files from the saves/ directory. 2. Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore embedded commands within the game data. 3. Capability inventory: The skill includes file-writing capabilities in scripts/game_engine.py and file-copying/importing logic in scripts/save_manager.py. 4. Sanitization: While the skill uses json.load to parse data, it does not sanitize or validate the resulting strings to prevent them from being interpreted as instructions by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 08:36 AM