homekit

Pass

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: The content consists of educational documentation and Swift boilerplate for standard iOS development using HomeKit and MatterSupport frameworks.\n- [DATA_EXFILTRATION]: The skill demonstrates access to home automation hierarchies (accessories, services, and characteristics). This data access is managed via standard iOS permission prompts (NSHomeKitUsageDescription) and is restricted to the application's legitimate context within the HomeKit database.\n- [PROMPT_INJECTION]: The skill documents how to ingest and display strings from the HomeKit environment (e.g., accessory.name), which presents a surface for indirect prompt injection. Developers should treat these user-controllable strings as untrusted if they are used to influence agent logic.\n
  • Ingestion points: SKILL.md and references/matter-commissioning.md (accessory.name, service.serviceType, characteristic.value)\n
  • Boundary markers: None present in the documentation.\n
  • Capability inventory: The skill enables reading and writing characteristic values (readValue, writeValue) and executing grouped actions (HMActionSet)\n
  • Sanitization: No explicit sanitization or validation of strings is shown in the provided Swift UI code samples.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 28, 2026, 08:14 PM