Agent Types

The Synari protocol supports five foundational agent archetypes. Each can be extended, mutated, or combined via prompt and config.

1. Autonomous Executor

Executes actions based on encoded strategies and environmental triggers. Ideal for agents that trade, rebalance, or initiate contract calls.

2. Observational Sentinel

Monitors data streams, APIs, or on-chain events to detect anomalies and issue intelligent responses or alerts.

3. Reasoning Analyst

Processes input, runs multi-step reasoning via LLM chains, and produces structured, explainable output. Useful for governance forecasting, simulations, or decision support.

4. Policy Enforcer

Evaluates other agents' actions against protocol rules. Can approve, reject, or escalate based on intent mismatch or constraint violations.

5. Custom Agent

A fully programmable agent using user-defined prompts, traits, and logic. Maximum flexibility for advanced developers and agent engineers.