skills/egorfedorov/slot-casino-game-developer-skills-for-stake-engine/ai-game-developer/Gen Agent Trust Hub
ai-game-developer
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script,
scripts/validate_ai_game_runtime.py, to perform structural and semantic validation on game runtime specifications. This is a core feature for verifying AI system integrations. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external specification files, creating a surface for indirect instructions to influence the agent.
- Ingestion points: The
scripts/validate_ai_game_runtime.pyscript reads user-provided JSON files (e.g.,ai_game_runtime_spec.json) via the--inputargument. - Boundary markers (absent): The prompt instructions do not specify any delimiters or warnings to isolate untrusted data within the JSON fields from the agent's instructions.
- Capability inventory: The skill relies on local script execution for validation and does not have network access or administrative permissions.
- Sanitization (absent): The validation logic focuses on data structure and types but does not perform sanitization of natural language fields like 'purpose' or 'name', which are subsequently analyzed by the agent.
Audit Metadata