Skip to main content

clawdesk-consensus

The consensus crate provides algorithms for multiple agents to reach agreement on shared decisions.

What It Does (Plain English)​

When multiple AI agents need to agree on something (like which answer is best, or what action to take), they need a fair voting system. This crate provides consensus mechanisms — ways for agents to vote, debate, and reach agreement. Think of it as a democratic decision-making process for AI agents.

Key Features​

  • PBFT (Practical Byzantine Fault Tolerance) — Agreement even if some agents misbehave
  • Voting — Weighted voting across agent opinions
  • Quorum — Configurable majority requirements
  • Conflict resolution — Handles disagreements between agents

Architecture Role​

LayerPosition
ServicesMulti-agent coordination

Dependencies​

  • clawdesk-types — Vote and decision type definitions
  • clawdesk-agents — Agent communication