EmblemAI — agentic threat model

Not certain from the listing — likely utilizes commercial LLMs to parse plain English commands into structured transaction payloads. Threats include prompt injection attacks that could trick the model into generating malicious transaction parameters or spoofing destination addresses.

Not certain from the listing — relies on real-time market data and blockchain state queries. Threats include data poisoning of market research feeds or RPC node spoofing, which could lead to incorrect balance displays or unfavorable trade executions.

Translates natural language commands into multi-chain smart contract interactions. Threats include insecure tool integration where the framework fails to validate the generated transaction parameters (e.g., slippage, recipient address) before presenting them to the user or executing them.

Not certain from the listing — requires secure handling of API keys and potentially wallet private keys or session states. Threats include server-side compromise leading to the theft of API credentials or unauthorized access to user wallet sessions.

Not certain from the listing — no mention of transaction simulation, pre-execution guardrails, or anomaly detection. Threats include a lack of observability into anomalous transaction volumes or suspicious destination addresses generated by the agent.

Not certain from the listing — lacks explicit details on multi-factor authentication, hardware wallet integration, or smart contract audits. Threats include weak authorization mechanisms allowing unauthorized users to trigger transactions via the assistant interface.

Not certain from the listing — primarily functions as a direct user-to-blockchain assistant. Threats include potential future integration with decentralized aggregators or third-party DeFi agents, introducing cascading trust and execution failures.