roblox-game-architecture
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from the Roblox DataModel and local scripts, creating a surface for indirect prompt injection where instructions embedded in project files could influence the agent.
- Ingestion points: The skill utilizes tools such as
script_read,script_grep,script_search,inspect_instance, andget_studio_stateto ingest project content (referenced in SKILL.md). - Boundary markers: The instructions do not define delimiters or specific warnings to prevent the agent from following instructions found within the analyzed scripts.
- Capability inventory: The skill's capabilities are limited to reading the environment state and scripts; no network access or file-writing tools are requested.
- Sanitization: No mechanisms for filtering or sanitizing the ingested script data are specified.
Audit Metadata