mirror of https://github.com/hanzoai/bot.git synced 2026-11-01 07:38:30 +00:00

Files

T

zeekay e5aac7b8a2 feat(skills): add complete Hanzo cloud service repertoire (20 new skills)

Full coverage of Hanzo infrastructure: MCP (260+ tools), CLI, LLM
Gateway, Chat, Network, Node, Web3, Consensus, ACI, REPL, Vector DB,
KV Store, PubSub, DocumentDB, Storage, Studio, Insights, Commerce,
Search, and Telemetry. Updated PaaS skill to use hanzo-tools-paas SDK.

2026-02-27 00:09:47 -08:00

2.4 KiB

Raw Permalink Blame History

name, description, metadata

name

description

metadata

hanzo-consensus

Metastable consensus protocol for multi-agent agreement. Two-phase finality with k-peer sampling for distributed AI decision making.

bot

requires

install

bins

python3

id	kind	package	label
pip	pip	hanzo-consensus	Install Hanzo Consensus (pip)

Hanzo Consensus — Multi-Agent Agreement

pip install hanzo-consensus

Metastable consensus protocol for distributed multi-agent decision making. Two-phase finality with configurable k-peer sampling.

Quick Start

from hanzo_consensus import Consensus, run

# Simple consensus among participants
result = run(
    participants=["agent-1", "agent-2", "agent-3"],
    proposal="Should we deploy to production?",
    rounds=10
)
print(result.decision)  # True/False
print(result.confidence)  # 0.0-1.0

Async Usage

from hanzo_consensus import Consensus

consensus = Consensus(
    rounds=10,
    k=3,           # Sample size per round
    alpha=0.8,     # Agreement threshold
    beta_1=15,     # Preference threshold (Phase I)
    beta_2=20      # Decision threshold (Phase II)
)

result = await consensus.run(participants, proposal)

MCP Mesh Consensus

from hanzo_consensus import MCPMesh, MCPAgent, run_mcp_consensus

# Create MCP-based consensus mesh
mesh = MCPMesh(agents=[
    MCPAgent("analyst", mcp_server="..."),
    MCPAgent("reviewer", mcp_server="..."),
    MCPAgent("approver", mcp_server="...")
])

result = await run_mcp_consensus(mesh, "Evaluate code quality of PR #42")

Protocol Parameters

Parameter	Description	Default
`rounds`	Number of sampling rounds	10
`k`	Peers sampled per round	3
`alpha`	Agreement threshold (0-1)	0.8
`beta_1`	Phase I preference threshold	15
`beta_2`	Phase II decision threshold	20

Protocol Design

Two-phase finality based on Lux Consensus:

Sampling Phase: Each round, sample k random peers and query their preference
Finality Phase: After sufficient consecutive agreements, finalize decision

Pure Python — no external dependencies.

2.4 KiB Raw Permalink Blame History