More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Lázaro Forwarder

    1

    Live inter-agent communication system: a transparent forwarder (kvproxy) + protocol for agents to wake each other up and talk back and forth without interrupting. Works with local LLMs (llama.cpp/router) and cloud APIs.

    $3.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Lázaro Forwarder

    Lázaro Forwarder

    Example session with this skill installed

    A local router at 127.0.0.1:8080 serving qwen36-mx. Point the forwarder at it, run a few agent sessions through it, then check /status to see KV cache effectiveness per turn.

    • Read your context and instructions
    • Compiled the l zaro forwarder

    Connects securely to your tools. The creator never sees your data.

    About this skill

    When you run multiple LLM sessions (local or cloud), they don't talk to each other by default. Each session is isolated. You need a human to copy-paste between them, or you lose context when one session dies.

    Lázaro Forwarder solves this with two pieces:

    1. kvproxy — a transparent forwarder that sits between your agent and the model endpoint. It forwards everything intact (including streaming), measures per-turn KV cache effectiveness (HIT/MISS), and exposes a /status endpoint. It works with any OpenAI-compatible endpoint: local router (:8080), cloud API, or anything in between.

    2. The Lázaro protocol — the rules for how agents talk to each other over that forwarder without stepping on each other. Wake the sleeping one, don't interrupt the awake one. Sign every message. Acknowledge before attending. Cap the chain at 3 hops.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.

    The protocol assumes OpenCode (or similar) as the agent runtime. The concepts transfer to any system with sessions, but the commands are OpenCode-specific. kvproxy is transparent by design — it does not modify requests or responses. If something breaks, it is not the forwarder. The protocol is a set of rules you adopt; the forwarder is infrastructure. You adapt both to your setup.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 15 days ago

    Software
    Python 3.10+OpenAI-compatible model endpoint

    Python needs only the standard library. The endpoint can be a local router such as llama-router or a cloud API.

    Permissions

    Allowed Hosts

    user-configured upstream endpoint (default: 127.0.0.1:8080)

    File Scopes

    none
    —
    the
    forwarder
    is
    in-memory
    only
    (last
    20
    turns)
    never
    writes
    to
    disk

    Purpose: transparent HTTP forwarder + KV cache effectiveness observer for inter-agent communication. Network: forwards requests to a single user-configured upstream endpoint; reads response bodies for metrics only. File I/O: none — in-memory only. Env vars: none required (all configuration via command-line flags). No eval(), no shell subprocesses, no curl|sh, no rm -rf, no base64 blobs. The forwarder is fully readable and commented.

    Listed15 days ago
    Updated7 days ago

    What's inside

    Frequently Asked Questions

    No. kvproxy is a transparent forwarder. It forwards everything intact (including streaming SSE) and only reads response bodies for metrics. If something breaks, it is not the forwarder.

    HIT means the KV cache is working — the model reuses cached attention states and processes fast (>3000 tok/s). MISS means the cache is not being reused — the model processes everything from scratch (<1200 tok/s). It tells you if your local LLM setup is efficient.

    Yes. Point --upstream at any OpenAI-compatible endpoint (OpenAI, DeepSeek, OpenRouter, etc.). The forwarder and protocol work the same way regardless of where the model runs.

    Because it is a system, not a tool. The Lázaro protocol (announcement, signature, acknowledgment, receiver discipline, hop limits) is intellectual property that you cannot find elsewhere. The forwarder is the infrastructure that makes it work. Together they solve a problem no one else has documented.