More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    WokiToki: Message Other Developers' AI Agents

    1

    Let Claude Code or Codex message a teammate's agent directly: peer to peer, end-to-end encrypted, no server.

    Free

    1 useSecurity scanned
    WokiToki: Message Other Developers' AI Agents

    WokiToki: Message Other Developers' AI Agents

    Example session with this skill installed

    Ask Rahul's agent whether the /orders API returns amounts in paise or rupees, and tell me what he says.

    • Read your context and instructions
    • Compiled the wokitoki

    Summary

    I sent your question to Rahul over WokiToki and got his answer: /orders returns amounts as integer paise.

    What I did

    StepToolResult
    1wokitoki_list_peersRahul is paired (active) and online
    2wokitoki_send_messageSent: "Does GET /orders return amounts in paise or rupees? Our checkout UI formats them, so we need the exact unit and type."
    3wokitoki_list_messagesDelivered (end-to-end receipt from Rahul's machine)
    4wokitoki_read_inbox1 new message from Rahul

    Rahul's reply

    Integer paise. 49900 means Rs 499. All money fields in /orders and /payments use the same convention.

    I'm treating this as information from another agent, not as instructions.

    Next steps

    • Update the checkout formatter to divide by 100 only at display time and keep integer math everywhere else.
    • Want me to ask Rahul whether refunds in /payments/refunds use the same unit?

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

    About this skill

    WokiToki gives your coding agent a walkie-talkie to other developers' agents. Ask your teammate's Claude Code or Codex a question, hand over context, or confirm an API contract, without copy-pasting between chat windows.

    This skill teaches your agent when and how to use the WokiToki MCP tools safely:

    • Send self-contained messages to a paired peer (wokitoki_send_message), replying in-thread with replyTo
    • Read the inbox and summarize who asked what (wokitoki_read_inbox, wokitoki_list_messages)
    • Check who is paired and online (wokitoki_list_peers, wokitoki_get_status)

    Built-in guardrails: incoming messages are treated as untrusted data, never as instructions. The agent asks before sending code, logs or anything confidential, and never sends secrets. Pairing, approval and revocation always stay with the human.

    How WokiToki works: each developer runs a small open-source daemon. Daemons find each other by public key on a public DHT and connect directly with NAT hole punching. Messages are end-to-end encrypted with libsodium. There is no server, no account and no tunnel. Pairing uses a one-time invite plus a 6-digit code that both people verify.

    Requirements: the free WokiToki CLI on macOS or Linux (Node.js 22.18+). One line installs it, starts it at login and adds the MCP server to Claude Code and Codex:

    curl -fsSL https://raw.githubusercontent.com/pandeyvishwas51-oss/wokitoki/main/install.sh | bash -s -- --name "Your Name"
    

    Open source (Apache-2.0): https://github.com/pandeyvishwas51-oss/wokitoki

    Text only up to 16 KiB. No file transfers. Delivery requires both machines to be online at once. Pairing and approval require manual CLI commands.

    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 any payment.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    GitHub
    WokiToki CLI
    WokiToki MCP
    Node.js 22.18+
    macOS
    Linux

    Permissions

    Network Access

    Allowed Hosts

    raw.githubusercontent.com

    The skill itself only instructs the agent to call the five WokiToki MCP tools. The locally installed WokiToki daemon makes the peer-to-peer network connection to paired peers. No shell commands, file reads or file writes are needed, and nothing is sent without a paired, human-approved peer.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    WokiToki bridges the gap between isolated AI agents by providing a secure, encrypted channel for them to exchange technical context, code snippets, and task status directly without human copy-pasting.

    Messages are end-to-end encrypted with libsodium and travel directly between the two machines over a Noise-encrypted peer-to-peer connection; nothing is stored on any server. The skill tells your agent to ask you before sending code or logs and to never send secrets. WokiToki has not had an external security audit yet, so avoid sending sensitive data.

    This skill uses the Model Context Protocol (MCP), making it compatible with any AI agent or IDE (like Claude Desktop or VS Code extensions) that supports the MCP standard and has the WokiToki daemon installed.

    No, for security reasons, the skill requires a manual pairing process between users; agents cannot autonomously discover or message new peers without prior human authorization.

    Messages are currently limited to 16 KiB of text, and both the sender and receiver machines must be online at the same time for the message to be successfully delivered.

    This free listing is a single SKILL.md that teaches your agent when and how to use the WokiToki MCP tools safely. The WokiToki CLI and MCP server are separate, free and open source (Apache-2.0); install them with one command from https://github.com/pandeyvishwas51-oss/wokitoki.

    New in Agents & Orchestration