roblox-cloud

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
  • [METADATA_POISONING]: The last_reviewed metadata field contains a future date (2026-08-07). While this is an anomaly, it does not introduce any security risks or deceptive behavior that would impact the safety of the skill.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes an attack surface where the agent interacts with external Roblox APIs and webhooks. The security posture is assessed as safe based on the following:
  • Ingestion points: Data enters the context via Roblox webhook notifications and user-initiated asset management tasks.
  • Boundary markers: The skill instructs the agent to 'let the user choose' and 'never assume permission,' maintaining human-in-the-loop control.
  • Capability inventory: The skill references external tools for asset acquisition (assets API, generate_mesh, generate_procedural_model), data persistence (data stores, ordered data stores, messaging), and ads management (Ads Manager API) but does not implement them itself.
  • Sanitization: Explicit instructions are provided to verify 'roblox-signature' headers, validate paths/schemas, and reject stale deliveries.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 06:23 AM
Security Audit — agent-trust-hub — roblox-cloud