networking

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is a collection of documentation, configuration, and reference code for Android networking. The provided Python script for validation uses safe loading practices, and the Kotlin snippets demonstrate proper security controls like certificate pinning.\n- [EXTERNAL_DOWNLOADS]: The skill references and includes configurations for libraries from trusted organizations, including Square (Retrofit, OkHttp, Moshi) and Google (GSON, AndroidX Security). These references are documented neutrally and follow secure implementation patterns.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes methods for processing untrusted data from external APIs.\n
  • Ingestion points: API responses processed via Retrofit UserService interfaces as shown in networking-apis.md and references/API_GUIDE.md.\n
  • Boundary markers: Use of structured JSON parsing with GSON and Moshi acts as a structural boundary.\n
  • Capability inventory: Includes network request capabilities (Retrofit/OkHttp) and local file writing (FileOutputStream) as part of its documented networking patterns.\n
  • Sanitization: Relies on type-safe model mapping and structural JSON validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 10:59 AM
Security Audit — agent-trust-hub — networking