Back to Ecosystem Pulse
ToolExperimental
K-LEAN
by calinfaja
Claude Code toolkit with multi-LLM consensus and specialist agents
Python
Updated Feb 16, 2026
Share:
Summary
Enhances Claude Code with a cross-platform toolkit that adds multi-LLM consensus, eight specialist agents, semantic knowledge search, and a one-command installer. Combines lightweight agent scaffolding, persistent memory, and consensus voting across LLMs to improve answer reliability and code review automation. Notable features include specialist agents for focused subtasks and integrations with lite LLM runtimes and router services for flexible backends.
Why It Matters
As agents are asked to make higher-stakes decisions, simple single-LLM outputs are no longer enough — you need reproducible signals and specialist workflows. K-LEAN brings multi-LLM consensus and specialist workflows into a Claude Code environment so teams can start capturing agent behavior, failure modes, and agreement signals without rebuilding orchestration from scratch. That makes it a practical starting point for experimenting with agent track record and early agent-to-agent evaluation patterns.
Best For
Practitioners prototyping multi-LLM consensus, code-review automation, and lightweight agent pipelines within Claude Code-based projects.
Applications
- Add multi-LLM consensus voting to code review and QA pipelines
- Prototype specialist agents (e.g., tester, reviewer, summarizer) that collaborate on a task
- Experiment with semantic memory and persistent knowledge search for agent workflows
Works With
anthropic
Topics
ai-agentsclaude-codeclaude-code-pluginclicode-review-automationk-leanlitellmnano-gptopenrouter-integrationpersistent-memory+1 more
Similar Tools
autogencrewai
Keywords
multi-agent trustagent-to-agent evaluationmulti-LLM consensusagent reliability