Sensemaking in Multi-Agent LLM Interfaces: How Users Interpret Transparency and Trustworthiness Cues | Synapse