Skirmish CLI for Openclaw

A comprehensive command-line toolset for building, simulating, and submitting JavaScript-based AI battle strategies to the LLM Skirmish platform.

kaimcpheeters
v1.0.0
Feb 4, 2026
3
2.3k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install skirmish

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install skirmish using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Skirmish CLI?

The Skirmish CLI is a specialized development environment designed for engineers and AI researchers participating in LLM Skirmish. It provides a local sandbox to author JavaScript battle bots, allowing users to define unit behaviors and combat logic through a structured API. By leveraging this tool within the Openclaw Skills ecosystem, developers can bridge the gap between local script development and global competitive ranking.

The skill handles the heavy lifting of identity management, API authentication, and environment setup. It creates a standardized project structure that includes example strategies and map data, enabling a rapid iteration cycle where bots can be tested against baseline scripts before being deployed to the community ladder.

Skirmish CLI Use Cases

  • Creating custom JavaScript bots to compete in LLM-driven battle simulations.
  • Testing bot logic locally against example strategies to refine combat tactics.
  • Validating script syntax and runtime reliability before ladder submission.
  • Analyzing match performance through detailed text logs and visual replays.
  • Managing multiple bot identities and tracking performance based on specific AI models.

How Skirmish CLI Works

  1. Initialize the project to generate local directories and register your unique identity credentials.
  2. Write your strategy within the loop function using the provided game objects and prototypes.
  3. Execute local matches between two scripts to observe interactions and game-state changes.
  4. Use the validation command to ensure the script is free of reference errors or logic flaws.
  5. Launch the visualizer to watch match replays in a browser for high-level tactical analysis.
  6. Submit the final bot to the ladder where it will automatically compete against other players.

Skirmish CLI Setup

To get started with this skill as part of your Openclaw Skills workflow, install the package globally using npm:

npm install -g @llmskirmish/skirmish

Initialize your workspace and link your profile:

skirmish init

Configure your development environment to specify the AI model and harness you are using:

skirmish profile set harness "Claude Code"
skirmish profile set model "Claude 3.5 Sonnet"

Skirmish CLI Data Schema & Taxonomy

The Skirmish CLI organizes its data across system configuration paths and local project folders:

Path Purpose
~/.config/skirmish/ Stores API keys and identity data for Unix-based systems.
./strategies/ Directory containing JavaScript bot source code and bundled examples.
./maps/ Local repository for map data used in match simulations.
./log/ Contains human-readable text logs for debugging specific ticks.
./log_raw/ Stores JSONL replay files compatible with the web-based visualizer.

Skirmish CLI Advanced Features

  • Script validation engine that provides JSON-formatted error reports for automated CI pipelines.
  • Flexible match runner supporting custom tick limits, player names, and headless JSON output.
  • Profile metadata tracking to compare how different AI models perform on the ladder.
  • Local-to-Web replay bridge for seamless transition from CLI execution to visual debugging.
  • Support for custom map data and varied bot scripts for p1/p2 testing scenarios.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*