hanzo-consensus 0.1.0

Metastable consensus protocol for multi-agent agreement

hanzo-consensus

Metastable consensus protocol for multi-agent agreement.

Install

pip install hanzo-consensus

Basic Usage

import asyncio
from hanzo_consensus import run, Result

async def execute(participant: str, prompt: str) -> Result:
    output = await call_agent(participant, prompt)
    return Result(id=participant, output=output, ok=True, ms=100)

async def main():
    state = await run(
        prompt="What is the best approach?",
        participants=["agent1", "agent2", "agent3"],
        execute=execute,
        rounds=3,
    )
    print(f"Winner: {state.winner}")
    print(f"Synthesis: {state.synthesis}")

asyncio.run(main())

MCP Mesh - Agent-to-Agent Consensus

Each agent in consensus is available as MCP to every other:

from hanzo_consensus import MCPMesh, run_mcp_consensus

# Create mesh of agents
mesh = MCPMesh()
mesh.register("claude", claude_server)
mesh.register("gpt4", gpt4_server)
mesh.register("gemini", gemini_server)

# Run 10 rounds of discussion
state = await run_mcp_consensus(
    mesh=mesh,
    prompt="Discuss the architecture",
    rounds=10,
)

# Access discussion history
for i, round_responses in enumerate(state.discussion_history):
    print(f"Round {i+1}:")
    for agent, response in round_responses.items():
        print(f"  [{agent}]: {response[:100]}...")

MCPMesh Features

• Agent Registration: Local FastMCP servers or remote endpoints
• Tool Calling: Any agent can call tools on any other agent
• Broadcasting: Call a tool on all agents simultaneously
• Discussion Rounds: Agents build on each other's responses

# Call tool on specific agent
result = await mesh.call("claude", "gpt4", "think", prompt="What do you think?")

# Broadcast to all agents
results = await mesh.broadcast("claude", "discuss", prompt="New proposal...")

Protocol

Two-phase finality:

Phase I (Sampling)

• Each participant proposes initial response
• k-peer sampling per round
• Luminance-weighted selection (faster = higher weight)
• Confidence accumulation toward β₁

Phase II (Finality)

• Threshold aggregation
• β₂ finality threshold
• Winner synthesis

Parameters

Param	Default	Description
rounds	3	Sampling rounds
k	3	Sample size per round
alpha	0.6	Agreement threshold
beta_1	0.5	Preference threshold (Phase I)
beta_2	0.8	Decision threshold (Phase II)

Reference

https://github.com/luxfi/consensus