hz-platform-sdk

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents APIs that handle untrusted data from the platform, such as deeplinks and session messages, which are common ingestion points for indirect instructions.\n
  • Ingestion points: The deeplinkMessage property in LaunchDetails (references/application-lifecycle.md) and GroupPresenceJoinIntent (references/group-presence.md) accepts opaque strings from external sources.\n
  • Boundary markers: Explicit boundary markers or instructions to ignore embedded commands are not included in the provided code snippets.\n
  • Capability inventory: The SDK documentation includes tools for launching other applications, downloading asset files, and performing network operations via Graph API access tokens.\n
  • Sanitization: Code examples demonstrate basic retrieval and navigation logic; input sanitization is expected to be handled by the application logic.\n- [EXTERNAL_DOWNLOADS]: The skill directs developers to download Maven artifacts from Meta's repository and Maven Central. These references are standard for software development and target trusted, well-known infrastructure for dependency management.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 09:15 AM