LaunchKit · 2026
Back to Skills

Multi-agent coordination, spatial memory, and swarm navigation.

0
760 downloads
by @a-sumo

Setup & Installation

openclaw skills install @a-sumo/eywa

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install eywa

What This Skill Does

Eywa is a coordination layer for multi-agent workflows. Agents connect to a shared room to see each other's activity, claim tasks, detect file conflicts, and store knowledge that persists across sessions.

Without shared state, parallel agents duplicate work, overwrite each other's changes, and lose learned context between sessions.

When to use it

  • Running parallel agents on a codebase without duplicate work
  • Detecting when two agents are about to edit the same file
  • Passing context from one agent session to the next
  • Tracking milestones across a multi-agent sprint
  • Storing discovered patterns so future sessions can reuse them

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: coordinate three agents to build auth, billing, and dashboard

AGENT
  1. 1Call eywa_start with a task description to get a room snapshot of active agents and queued tasks
  2. 2Call eywa_tasks to see available work, then eywa_pick_task to claim the auth task
  3. 3Call eywa_claim to declare file scope and trigger conflict detection with other agents
  4. 4Log each significant operation with eywa_log, tagging system, action, and outcome
  5. 5Call eywa_learn to store reusable patterns, then eywa_done with a summary and artifact list
OUTPUT

All three agents complete their tasks without conflicts, with shared knowledge and progress visible across the room

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

eywa-ai npm package (provides eywa CLI)EYWA_ROOM environment variableEYWA_AGENT environment variable