threejs-game-director
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local scripts to facilitate the development workflow and ensure quality standards. The
scripts/probe_asset_credentials.shshell script is used to check the user's environment profiles (e.g.,.bashrc,.zshrc) for the presence of specific API keys (TRIPO_API_KEY,GEMINI_API_KEY,ELEVENLABS_API_KEY). This script safely reports only the status (SET or MISSING) and does not output the secret values themselves. Additionally, thescripts/audit_reference_report.pyPython script performs static analysis on local markdown reports to verify the completion of required project phases and ledgers. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and process data from external AI generators and user-provided reports.
- Ingestion points: The agent ingests Task IDs, external 3D models (GLB/GLTF/FBX), and 2D image assets from third-party AI services. It also processes local markdown reports during the audit phase.
- Boundary markers: The instructions do not specify the use of clear delimiters or warnings to ignore potential instructions embedded within external asset metadata or generated reports.
- Capability inventory: The skill has capabilities to modify the file system (creating project structures) and execute local scripts using Python, Node.js, and Bash.
- Sanitization: There is no evidence of specific sanitization or validation logic to filter or escape content from external assets before they are utilized by the agent or incorporated into the Three.js game code.
Audit Metadata